RU2426180C2 - Calculation and adjustment of audio signal audible volume and/or spectral balance - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: audio signal processing makes sense in volume adjustment with volume compensation, dynamic correction and compensation of background noise in audio playback hardware. Modification parameters are used to modify audio signal to reduce difference between specific volume and target specific volume.

EFFECT: higher intelligibility.

26 cl, 19 dwg

Description

FIELD OF TECHNOLOGY

The invention relates to the processing of audio signals. More specifically, the invention relates to measuring and adjusting the perceived sound volume and / or the perceived spectral balance of an audio signal. The invention, for example, is useful in one or more of: volume control with volume compensation, automatic gain control, dynamic range adjustment (including, for example, limiters, compressors, dynamic range expanders, etc.), dynamic correction and compensation noise background interference in audio playback environments. The invention includes not only methods, but also related computer programs and apparatus.

BACKGROUND

There have been many attempts to develop a satisfactory objective way to measure volume. Fletcher and Munson determined in 1933 that human hearing is less sensitive at low and high frequencies than at medium (or voice) frequencies. They also found that the relative change in sensitivity decreased as the sound level increased. The previous volume meter consisted of a microphone, an amplifier, a meter and a combination of filters designed to roughly copy the frequency response of hearing at low, medium and high sound levels.

Even if such devices provided measurements of the volume of a single isolated tone of a constant level, measurements of more complex sounds did not very well correspond to the subjective sensations of volume. Sound level meters of this type were standardized, but were used only for specific tasks, such as dosimetric monitoring and industrial noise surveillance.

In the early 1950s, Zwicker and Stevens, among other things, continued the work of Fletcher and Munson to develop a more realistic model of the process of perceiving loudness. Stevens published a method for “Calculating the Volume of Mixed Noise” in the journal of the Acoustic Society of America in 1956, and Zwicker published his article “Psychological and Methodical Basis of Loudness” in Acoustica in 1958. In 1959, Zwicker published a graphical method for calculating volume, as well as several similar articles shortly thereafter. Stevens and Zwicker methods were standardized as ISO 532, parts A and B (respectively). Both methods included similar steps.

First of all, the time-dependent energy distribution along the basilar membrane of the inner ear, referred to by reference as pumping, is simulated by the passage of an audio signal through a comb of band-pass auditory filters with central frequencies uniformly spaced along the step scale of critical bands. Each auditory filter is designed to simulate a frequency response at a specific location along the basilar membrane of the inner ear, with a central filter frequency corresponding to that location. The critical bandwidth is defined as the bandwidth of one such filter. Measured in Hertz units, the width of the critical band of these auditory filters increases with increasing center frequency. Therefore, it is useful to determine the curvilinear frequency scale from the condition that the width of the critical band for all auditory filters, measured on this curvilinear scale, is constant. Such a curved scale is indicated by reference as a stepped scale of critical bands and is very useful in understanding and simulating a wide range of physiological phenomena. For example, look Psychoacoustics - Facts and Models by E. Zwicker and H. Fasti, Springer-Verlag, Berlin, 1990 (Psychoacoustics - Facts and Models by E. Zvikeru and H.Fasti, Springer-Verlag, Berlin, 1990). The Stevens and Svicker methods use a step scale of critical bands, referred to as the Bark scale, in which the width of the critical band is constant below 500 Hz and increases above 500 Hz. Moore and Glazberg later defined a stepped critical band scale, which they called the rectangular equivalent bandwidth (ERB) scale (BCJ Moore, B. Glasberg, T. Baer, “A Model for the Prediction of Thresholds, Loudness, and Partial Loudness” , Journal of the Audio Engineering Society, Vol. 45, No. 4, April 1997, pp. 224-240 (B. C. J. Moore, B. Glasberg, T. Baer, “Model for predicting thresholds, volume and of partial tones ”, Journal of the Sound Engineering Community, Volume 45, No. 4, April 1997, pp. 224-240)). Through psychoacoustic experiments using narrow-band gap noise maskers in the spectrum, Moore and Glazberg demonstrated that the width of the critical band continues to decrease below 500 Hz, in contrast to the Bark scale, where the width of the critical band remains constant.

The subsequent calculation of the pump is a non-linear compression function, which forms a parameter referred to by the link as "specific volume". Specific loudness is a measure of perception loudness as a function of frequency and time and can be measured in units of perception loudness per unit frequency on a step scale of critical bands, such as the Bark or ERB scale discussed above. Specifically, the specific volume represents a continuous distribution of volume as a function of frequency and time, and the time-dependent “full volume” is calculated by integrating this frequency distribution. In practice, an accurate perception of specific loudness is obtained by discretizing this distribution uniformly on the step scale of critical bands, for example, by using the auditory filters mentioned above. In this case, the total volume can be calculated by simply summing the specific volume from each filter. To reduce complexity, some applications can calculate a rough approximation for specific volume due to minor inaccuracies in the estimation and modification of perceived volume. Such approximations will be discussed in more detail later.

Volume can be measured in units of backgrounds. The volume of the sound specified in the background is the sound pressure level (SPL) of the tone at 1 kHz, which has a subjective volume equal to that of the sound. Traditionally, the 0 dB reference point for the SPL is the rms pressure of 2 × 10 ^-5 Pascals, and therefore this is also the reference point of 0 backgrounds. Using this definition when comparing the volume of tones at frequencies other than 1 kHz with the volume of 1 kHz, an equal volume curve can be determined for the level set in the backgrounds. 11 shows curves of equal loudness for frequencies between 20 Hz and 12.5 kHz, and for levels in the backgrounds between 4.2 backgrounds (considered the threshold of audibility) and 120 backgrounds (ISO226: 1087 (E), “Acoustics - Normal equal loudness level contours ”(“ Acoustics - normal curves of equal volume level ”)). The measurement in the background takes into account the changing sensitivity of human hearing depending on the frequency, but the results do not provide an opportunity to estimate the relative subjective sound volumes at variable levels, since there is no attempt to introduce a correction for the non-linearity of the increase in volume depending on the SPL, that is, for the fact that the interval the curves are changing.

Volume can also be measured in units of "sleep". There is an unambiguous correspondence between units of backgrounds and units of sleep, which is indicated in Fig. 11. One dream is defined as the volume of a 1 kHz unmodulated harmonic wave at 40 dB (SPL) and equals 40 backgrounds. The units of sleep are such that a double increase in sleep corresponds to a doubling of perceived loudness. For example, 4 sleeps are perceived as a volume twice as large as that of 2 sleeps. Thus, the expression of volume levels in sleep is more informative. Provided that specific loudness is determined as an indicator of perception loudness as a function of frequency and time, specific loudness can be measured in units of sones per unit frequency. Thus, when using the Bark scale, the specific volume has units of sleeps per Bark and, similarly, using the ERB scale, units are sones per ERB.

As mentioned above, the sensitivity of the human ear varies both with frequency and level, a circumstance well documented in the literature on psychoacoustics. One of the consequences is that the perceived spectrum or timbre of a given sound changes depending on the acoustic level at which the sound is heard. For example, for sound containing low, medium, and high frequencies, the perceived relative proportions of such frequency components change with the overall sound volume; when it is quiet, the low and high frequency components sound quieter relative to the mid frequencies than they sound when it is loud. This phenomenon is well known, and has been reduced in sound reproduction equipment through the so-called loudness volume controls. A loudly-controlled volume control is a volume control that applies a low-frequency, and sometimes also a high-frequency boost, as the volume decreases. Thus, the lower sensitivity of the ear at extreme frequencies is compensated by the artificial rise of such frequencies. Such regulators are completely passive; the degree of compensation applied is a function of adjusting the volume control or some other user-controlled control, not as a function of the content of the audio signals.

In practice, changes in the perceived relative spectral balance between low, medium and high frequencies depend on the signal, in particular on its effective spectrum and on whether it is intended to be loud or quiet. Consider a recording of a symphony orchestra. Played at the same level as a member of the public would hear in a concert hall, the balance across the width of the spectrum can be correct, the orchestra plays loudly or quietly. If music is played, for example, 10 dB quieter, the perceived balance across the width of the spectrum changes in one way for loud passages and changes in another way for quiet passages. The traditional passive loudness volume control does not apply different compensations as a function of music.

In international patent application No. PCT / US 2004/016964, registered May 27, 2004, published December 23, 2004 as WO 2004/111994 A2, Schifeldt and others disclose, inter alia, a system for measuring and adjusting the perceived volume of an audio signal. The PCT application, which indicates the United States, is hereby incorporated by reference in its entirety. In the said application, the psychoacoustic model calculates the volume of the sound signal in units of perception. In addition, the application establishes a technology for calculating a broadband multiplicative gain, which, when applied to audio, results in a loudness of the gain-modified audio essentially being the same as the reference loudness. However, the use of such broadband amplification changes the perceived spectral balance of audio.

SUMMARY OF THE INVENTION

In one aspect, the invention provides for extracting information used to control the specific volume of an audio signal by modifying the audio signal in order to reduce the difference between its specific volume and target specific volume. Specific volume is a measure of perception volume as a function of frequency and time. In practical implementations, the specific volume of the modified audio signal can be made approaching the target specific volume. The approximation may be influenced not only by considerations of conventional signal processing, but also by temporal and / or frequency smoothing, which can be applied when modifying, as described below.

Since specific loudness is a measure of the loudness of perception of an audio signal as a function of frequency and time, in order to reduce the difference between the specific loudness of the audio signal and the target specific loudness, the modification can modify the audio signal as a function of frequency. Although in some cases the target specific loudness may be time-independent, and the sound signal itself may be a steady-state time-independent signal, typically the modification may also modify the sound signal as a function of time.

Aspects of the present invention can also be used to compensate for background noise interfering with the audio playback environment. When audio is heard in the presence of background noise, the noise can partially or completely mask the audio in some way, depending both on the level and spectrum of the audio, and on the level and spectrum of the noise. The result is a restructuring of the perceived audio spectrum. According to psychoacoustic teaching (e.g., see Moore, Glasberg, and Baer, “A Model for the Prediction of Thresholds, Loudness, and Partial Loudness”, J. Audio Eng. Soc, Vol. 45, No. 4, April 1997 ( Moore, Glazberg and Baer, “A Model for Predicting Thresholds, Partial Tone Loudness and Volume”, Journal of the Audio Engineering Community, Volume 45, No. 4, April 1997)), one can define the “specific volume of partial tones” of an audio as the audio perception loudness in the presence of a secondary interfering sound signal, such as noise.

Thus, in yet another aspect, the invention provides for extracting information used to control the specific volume of an audio signal by modifying the audio signal in order to reduce the difference between its specific volume of partial tones and the target specific volume. Doing this mitigates the effects of noise in a sensible, precise way. In this and other aspects of the invention, which take into account the interfering noise signal, it is assumed that there is access separately to the audio signal and separately to the secondary interfering signal.

In yet another aspect, the invention provides for controlling the specific volume of the audio signal by modifying the audio signal in order to reduce the difference between its specific volume and the target specific volume.

In yet another aspect, the invention provides for controlling the specific volume of partial tones of an audio signal by modifying the audio signal in order to reduce the difference between its specific volume and target specific volume.

When the target specific volume is not a function of the audio signal, it may be a stored and received target specific volume. When the target specific volume is not a function of the audio signal, modifying or receiving can explicitly or implicitly calculate the specific volume or specific volume of the partial tones. Examples of implicit calculations include a look-up table or “reflective series of solutions” mathematical expression in which specific loudness and / or specific loudness of partial tones is determined by nature (a term reflecting a series of solutions is mentioned to describe a mathematical expression that can be accurately represented using a finite number of standard mathematical operations and functions, such as exponentiation and cosine). In addition, when the target specific volume is not a function of the audio signal, the target specific volume may be independent of both time and frequency, or it may be independent of time only.

In yet another aspect, the invention provides for processing an audio signal by processing an audio signal or an indicator of an audio signal in accordance with one or more process sequences or one or more process control parameters to form a target specific volume. Although the target specific volume may be time-independent (“constant”), the target specific volume may advantageously be a function of the specific volume of the audio signal. Although it may be a static, frequency and time independent signal, typically the audio signal itself is frequency and time dependent, thereby causing the target specific volume to be frequency and time dependent when it is a function of the audio signal.

The audio and target specific volume or representation of the target specific volume may be received from the transmitted data or reproduced from the storage medium.

Introduction of the target specific loudness may be one or more scale factors that scale the audio signal or the sound signal component.

The target specific volume of any of the above aspects of the invention may be a function of an audio signal or an indicator of the audio signal. One suitable indicator of an audio signal is the specific volume of the audio signal. The function of an audio signal or an indicator of an audio signal may be a scaling of an audio signal or an indicator of an audio signal. For example, scaling can be one or a combination of scaling:

(a) time and frequency-dependent scale factor Ξ [ b , t ] scaling the specific volume, as a function of

(b) time-dependent, frequency-independent frequency coefficient Φ [ t ] scaling the specific volume, as a function of

(c) time-independent, frequency-dependent scale factor Θ [ b ] scaling the specific volume, as a function of

(d) time-independent, frequency-independent frequency coefficient α scaling the specific volume of the audio signal, as a function of

[b, t] - целевая удельная громкость, N[b, t] - удельная громкость звукового сигнала, b - показатель частоты, а t - показатель времени.in which

[ b , t ] is the target specific volume, N [ b , t ] is the specific volume of the audio signal, b is a frequency indicator, and t is a time indicator.

In the case (a) of a time and frequency dependent scaling factor, scaling can be determined at least in part by the ratio of the desired multiband volume and multiband volume of the audio signal. Such scaling can be used as a dynamic range control. Additional details of aspects of the application of the invention as a dynamic range controller are set forth below.

In addition, in case (a) of a time-frequency and frequency-dependent scale factor, the specific loudness can be scaled by the ratio of the index of the desired spectral shape to the spectral shape of the sound signal. Such scaling can be used to convert the perceived spectrum of an audio signal from a time-dependent perceived spectrum to a substantially time-independent perceived spectrum. When the specific volume is scaled by the ratio of the index of the desired spectral shape to the spectral shape of the sound signal, such scaling can be used as a dynamic equalizer. Further details of aspects of applying the invention as a dynamic equalizer are set forth below.

In the case (b) of a time-dependent frequency-independent scale factor, scaling can be determined at least in part by the ratio of the desired broadband volume and broadband volume of the audio signal. Such scaling can be used as an automatic gain control or dynamic range control. Further details of the aspects of applying the invention as an automatic gain control and dynamic range controller are set forth below.

In the case of (a) (time-dependent and frequency-dependent scale factor) or case (b) (time-dependent, independent of the frequency of the scale factor), the scale factor may be a function of the sound signal or an indicator of the sound signal.

In both, the case (c) of the time-independent frequency-dependent scale factor or the case (d) of the time-independent frequency-dependent scale factor, the modification or receipt may include storing the scale factor, or the scale factor may be received from an external source .

In either case (c) and (d), the scaling factor may not be a function of the sound signal or the sound signal metric.

In any of the various aspects of the invention and its variants, the modification, production, or formation can differently explicitly or implicitly calculate (1) the specific volume and / or (2) the specific volume of partial tones, and / or (3) the target specific volume. Implicit calculations, for example, can include a look-up table or a mathematical expression that reflects a number of solutions.

Modification parameters can be smoothed over time. Modification parameters, for example, can be (1) a set of amplitude scaling factors related to the frequency bands of an audio signal, or (2) a set of filter coefficients for controlling one or more filters, such as a multi-tap (with a finite impulse response, FIR) FIR filter or multipolar (with infinite impulse response, IIR) IIR filter. The scaling factors or filter coefficients (and the filters to which they are applied) may be time dependent.

When calculating the function of the specific loudness of an audio signal, which determines the target specific loudness, or the inversions of such a function, the sequence of operations or the sequence of operations performing such calculations work in what can be described as a region of (psychoacoustic) loudness of perception - by input data and output data calculation are specific volume. In contrast, when applying amplitude scaling factors to frequency bands of an audio signal or applying filter coefficients to adjustable filtering of an audio signal, the modification parameters act to modify the audio signal outside the region of (psychoacoustic) loudness of perception, in what can be characterized as the region of electrical signals. Although modifications to the sound signal can be made with respect to the sound signal in the field of electrical signals, such changes in the field of electrical signals are obtained from calculations in the (psycho-acoustic) region of perception loudness, provided that the modified sound signal has a specific loudness that approaches the desired target specific volume.

By obtaining the modification parameters from calculations in the loudness region, greater control over the loudness of perception and the spectral balance of perception can be achieved than if such modification parameters were obtained in the field of electrical signals. In addition, the use of a psychoacoustic filter comb to simulate a basilar membrane or its equivalents when performing calculations in the volume domain can provide more detailed control of the perceived spectrum than in layouts that receive modification parameters in the field of electrical signals.

Each of the modification, production, and generation may be dependent on one or more of an interfering audio signal indicator, target specific volume, an estimate of the specific volume of the unmodified audio signal obtained from the specific volume or specific volume of partial tones of the modified audio signal, the specific volume of the unmodified audio signal, and approximations to the target specific volume obtained from the specific volume or specific volume of the partial tones of the modified sound signal.

Modification or acquisition may obtain modification parameters at least partially from one or more of an interfering sound signal indicator, target specific volume, estimated specific volume of an unmodified audio signal, obtained from specific volume or specific volume of partial tones of a modified audio signal, specific volume of an unmodified audio signal and approaching the target specific volume obtained from the specific volume or specific volume of partial tones differential sound signal.

More specifically, the modification or preparation can obtain the modification parameters at least partially from

(1) one of

target specific volume, and

estimating the specific volume of the unmodified audio signal obtained from the specific volume of the modified audio signal, and

(2) one of

the specific volume of the unmodified sound signal, and

approaching the target specific volume obtained from the specific volume of the modified audio signal,

or, when an interfering audio signal has to be taken into account, the modification or acquisition may obtain the modification parameters at least partially from

(1) an indicator of an interfering sound signal,

(2) one of

target specific volume, and

estimating the specific volume of the unmodified audio signal obtained from the specific volume of partial tones of the modified audio signal, and

(3) one of

the specific volume of the unmodified sound signal, and

approximations to the target specific volume obtained from the specific volume of partial tones of the modified audio signal.

A direct coupled arrangement may be used in which the specific volume is obtained from an audio signal and in which the target specific volume is received from a source external to the method, or from storage when the modification or acquisition includes storage of the target specific volume. Alternatively, a mixed direct feedback / feedback arrangement may be used in which the approximation of the target specific volume is obtained from a modified sound signal, and in which the target specific volume is received from a source external to the method, or from storage when the modification or receipt includes storing the target specific volume.

Modification or receipt may include one or more sequences of operations for obtaining, explicitly or implicitly, the target specific volume, one or those of which calculate, explicitly or implicitly, the function of the sound signal or indicator of the sound signal. In one alternative, a direct-coupled arrangement may be used in which the specific volume and target specific volume are obtained from an audio signal, and obtaining the target specific volume uses the function of an audio signal or an indicator of the audio signal. In one alternative, a mixed direct feedback / feedback arrangement may be used in which the approximate target specific loudness is obtained from a modified audio signal and the target specific loudness is obtained from an audio signal, obtaining the target specific loudness uses the function of an audio signal or an indicator of an audio signal .

Modification or receipt may include one or more sequences of operations for obtaining, explicitly or implicitly, an estimate of the specific volume of an unmodified audio signal in response to a modified audio signal, either of which, explicitly or implicitly, calculate the inverse of the function of the audio signal or sound indicator signal. In one of the alternatives, a feedback arrangement is used in which an estimate of the specific volume of an unmodified sound signal and an approximation to a target specific volume are obtained from a modified sound signal, an estimate of the specific volume is calculated using the inverse of the function of the audio signal or an indicator of the audio signal. In yet another alternative embodiment, a mixed direct feedback / feedback arrangement is applied in which the specific volume is obtained from an audio signal and the specific volume estimate of an unmodified audio signal is obtained from a modified audio signal, an estimate is calculated using an inversion of said audio signal function or sound indicator signal.

Modification parameters can be applied to the audio signal to form a modified audio signal.

Another aspect of the invention is that there may be a temporal and / or spatial separation of the sequences of operations or devices, so that, in fact, there is an encoder or coding, as well as a decoder or decoding. For example, there may be an encoding / decoding system in which a modification or reception can transmit and receive or store, as well as reproduce an audio signal and either (1) the modification parameters or (2) the target specific volume or representation of the target specific volume. As an alternative, in reality, there can only be an encoder or coding in which there is a transmission or storage of an audio signal and (1) modification parameters, or (2) a target specific volume or a representation of a target specific volume. As an alternative, as mentioned above, in reality, there can only be a decoder or decoding in which there is a reception or reproduction of an audio signal and (1) modification parameters or (2) target specific volume or representation of target specific volume.

DESCRIPTION OF DRAWINGS

Figure 1 is a functional block diagram illustrating an example implementation with direct connection according to aspects of the invention.

2 is a functional block diagram illustrating an example implementation with feedback according to aspects of the invention.

3 is a functional block diagram illustrating an implementation example with mixed direct feedback / feedback according to aspects of the invention.

4 is a functional block diagram illustrating an example of another implementation with mixed direct feedback / feedback according to aspects of the invention.

5 is a functional block diagram illustrating the manner in which an unmodified audio signal and modification parameters that are defined by any one of the direct link, feedback, or mixed direct / link configurations can be stored or transmitted for use, for example, in a temporally and spatially separated arrangement or sequence of operations.

6 is a functional block diagram illustrating the manner in which the unmodified audio signal and the target specific volume or its presentation, which are defined by any one of the direct communication, feedback or mixed direct / mixed arrangements, can be stored or transmitted for use, for example, in a temporally and spatially separated device or sequence of operations.

7 is a schematic functional block diagram or a schematic flowchart of a method showing an overview of an aspect of the present invention.

Fig. 8 is an idealized typical characteristic of a linear filter P ( z ) suitable as a transmission filter in an embodiment of the present invention, in which the vertical axis is decibel attenuation (dB) and the horizontal axis is the logarithmic base 10 frequency in Hertz (Hz).

Figure 9 shows the relationship between the frequency scale ERB (vertical axis) and frequency in Hertz (horizontal axis).

10 shows a set of idealized auditory filter sample characteristics that approximate the definition of a critical band on an ERB scale. The horizontal scale is the frequency in Hertz, and the vertical scale is the level in decibels.

11 shows curves of equal loudness according to ISO 226. The horizontal scale is the frequency in Hertz (logarithmic, base 10, scale), and the vertical scale is the sound pressure level in decibels.

12 shows equal volume curves according to ISO 226 normalized by the transmission filter P (z). The horizontal scale is the frequency in Hertz (a logarithmic base 10 scale), and the vertical scale is the sound pressure level in decibels.

13a is an idealized graph showing broadband and multiband gain factors for scaling a volume of 0.25 on a female speech segment. The horizontal scale is the ERB bands, and the vertical scale is the relative decibel gain (dB).

Fig.13b is an idealized graph showing the specific loudness, respectively, of the original signal modified by the broadband signal gain and modified by the multiband signal gain. The horizontal scale is the ERB bands, and the vertical scale is the specific volume (sleep / ERB).

Fig. 14a is an idealized graph showing: L _o [ t ] as a function of L _i [ t ] for a typical AGC. The horizontal scale is log ( L _i [ t ]), and the vertical scale is log ( L _o [ t ]).

14b is an idealized graph showing: L _o [ t ] as a function of L _i [ t ] for a typical DRC. The horizontal scale is log ( L _i [ t ]), and the vertical scale is log ( L _o [ t ]).

15 is an idealized graph showing a typical band smoothing function for a multi-band DRC. The horizontal bar is the band number, and the vertical bar is the gain output for band b.

FIG. 16 is a schematic functional block diagram or schematic flowchart of a method showing an overview of an aspect of the present invention.

FIG. 17 is a schematic functional block diagram or schematic flowchart of a method similar to FIG. 1, which also includes noise compensation in a reproducing environment.

BEST MODE FOR CARRYING OUT THE INVENTION

Figs 1 to 4 show functional block diagrams illustrating possible examples of direct feedback feedback implementations and two mixed direct feedback / feedback embodiments according to aspects of the invention.

With reference to an example topology with a direct connection in FIG. 1, an audio signal is fed into two paths: (1) a signal path containing a sequence of operations, or a device 2 (“Modify an audio signal”) capable of modifying audio in response to modification parameters, (2) a control path containing a control sequence or device 4 (“Generate modification parameters”) capable of generating such modification parameters. To modify the audio signal 2 in the example of the direct link topology of FIG. 1 and in each of the examples of FIGS. 2-4 it is possible with a device or a sequence of operations that modify the audio signal, for example, its amplitude, depending on the frequency and / or time in accordance with the parameters M modifications adopted from Generate modification parameters, 4, (or from equivalent sequences of operations or devices 4 ', 4 "and 4' '' in each of the examples of Figures 2-4, respectively). Generate modification parameters, 4, and its equivalents in figure 2-4 each Modify the audio signal, 2, operates in the field of electrical signals and generates a modified audio signal in each of the examples of figures 1-4. In addition, in each of the examples of figures 1-4, Modify the audio signal, 2, and Generate modification parameters, 4, (or its equivalents) modify the audio signal to reduce the difference between its specific volume and target specific volume.

In the direct-link example of FIG. 1, the flowchart or device 4 may include several flowcharts and / or devices: a flowchart or device 6 “Calculate target specific volume”, which calculate the target specific volume in response to an audio signal or an audio parameter a signal, such as a specific volume of an audio signal, a sequence of operations or a device 8 "Calculate specific volume", which calculate the specific volume of an audio signal in m to audio signal or measure of audio signals, such as its pump and the flow device 10 or "Calculate modification parameters", which are calculated modification parameters in response to the specific loudness and target specific loudness. Calculate the target specific volume, 6, can perform one or more functions " F ", each of which may have the parameters of the functions. For example, it can calculate the specific loudness of an audio signal, and then apply one or more of the F functions to it to provide the target specific loudness. This is schematically shown in FIG. 1 as an input signal, “Select function (s) F and parameter (s) of function (s)” in a sequence of operations or device 6. Instead of calculating by the device or sequence of 6 operations, the target specific volume may be issued by a sequence of operations or a storage device (schematically shown as an input signal "Stored" in the sequence of operations or device 10) included in or associated with the Generate modification parameters, 4, or source, external relative to the whole of the flowchart or device (shown schematically as the input signal "External" in the flowchart or device 10). Thus, the modification parameters are based, at least in part, on calculations in the area of (psychoacoustic) loudness of perception (i.e., at least specific loudness and, in some cases, calculations of the target specific loudness).

The calculations performed by the sequences of operations or devices 6, 8 and 10 (and the sequences of operations or devices 12, 14, 10 'in the example of FIGS. 2, 6, 14, 10' 'in the example of FIG. 3, and 8, 12, 10' '' in the example of FIG. 4) can be performed explicitly and / or implicitly. Examples of explicit execution include (1) a look-up table whose records are based, in whole or in part, on the specific volume and / or target specific volume and / or calculation of the modification parameters, and (2) a mathematical expression reflecting a series of solutions, which, according to by its nature, based, in whole or in part, on the specific loudness and / or target specific loudness, and / or modification parameters.

Although the sequence of operations or device 6, 8 and 10 of the calculation according to the example of figure 1 (and the sequence of operations or device 12, 14, 10 'in the example of figure 2, 6, 14, 10' 'in the example of figure 3, and 8, 12, 10 ″ ″ in the example of FIG. 4) are schematically shown and described as separate, this is intended for purposes of explanation only. It will be understood that those or all of these sequences of operations or devices can be combined into single sequences of operations or devices or combined in different ways in multiple sequences of operations or devices. For example, in the arrangement of FIG. 9 below, a direct-link topology, as in the example of FIG. 1, a process or device that calculates modification parameters does this in response to a smoothed pump obtained from the audio signal and the target specific volume. In the example of Fig. 9, a device or a sequence of operations that calculate modification parameters implicitly calculates the specific volume of an audio signal.

[b, t]) может рассчитываться масштабированием удельной громкости (N [b, t]) одним или более коэффициентами масштабирования. Масштабирование может быть зависящим от времени и частоты масштабным коэффициентом Ξ[b, t], масштабирующим удельную громкость, как в зависимостиAs an aspect of the present invention, in the example of FIG. 1 and in other examples of embodiments of the invention provided herein, the target specific volume (

[ b , t ]) can be calculated by scaling the specific volume ( N [ b, t ]) with one or more scaling factors. Scaling can be a time-frequency-dependent scale factor Ξ [ b , t ] scaling the specific volume, as a function of

time-dependent, frequency-independent scale factor Φ [ t ] scaling the specific volume, as a function of

time-independent, frequency-dependent scale factor Θ [ b ] scaling the specific volume, as a function of

scale factor α , scaling the specific volume of the sound signal, as depending

where b is a measure of frequency (e.g., band number), and t is a measure of time (e.g., frame number). Numerous scaling may also be applied using multiple instances of a particular scaling and / or combinations of specific scaling. Examples of such multiple scaling are given below. In some cases, as further explained below, scaling may be a function of the audio signal or indicators of the audio signal. In other cases, as also further explained below, when scaling is not a function of an indicator of the audio signal, scaling may be determined or otherwise provided. For example, the user could select or apply scaling with a time and frequency independent scale factor α or with a time independent frequency dependent frequency scale factor Θ [ b ].

Thus, the target specific volume can be expressed as one or more functions F of the audio signal or indicator of the audio signal (specific volume is one of the possible indicators of the audio signal):

[b, t]):Provided that the function or functions F are reversible, the specific volume ( N [ b , t ]) of the unmodified sound signal can be calculated as the inverse function or functions F ^{-1 of the} target specific volume (

[ b , t ]):

As will be seen below, the inverse function or functions F ^{−1 are} calculated in the examples with feedback and mixed direct feedback / feedback in FIGS. 2 and 4.

The input “Select function (s) and function parameter (s)” for Calculate target specific loudness, 6, is shown to indicate that a device or process sequence 6 can calculate target specific loudness by applying one or more functions in accordance with one or more feature options. For example, Calculate the target specific volume, 8, can calculate the function or functions “ F ” of the specific volume of the audio signal in order to determine the target specific volume. For example, the input signal “Select function (s) and function parameter (s)” may select one or more specific functions that fall into one or more of the above types of scaling along with one or more parameters of functions, such as constants (for example, scale factors ) related to functions.

The scaling factors associated with scaling can serve as a representation of the target specific loudness since the target specific loudness can be calculated as the specific specific loudness scaling, as indicated above. Thus, in the example of FIG. 9 described below and mentioned above, the look-up table can be indexed by scaling factors and pumpings so that the calculation of the specific volume and the target specific volume is inherent in the table.

Using a lookup table that reflects a series of mathematical expression solutions or some other technology, the operation Generate modification parameters, 4, (and its equivalent sequences of operations or devices 4 ', 4 "and 4' '' in each of the examples of Figures 2-4) is such that the calculations are based on the region of (psychoacoustic) loudness of perception, even if the specific loudness and target specific loudness may not be calculated explicitly.There is an explicit specific loudness or there is speculative, implicit specific loudness. Thus, there is an explicit target specific loudness or there is a speculative, implicit target specific loudness.In any case, the calculation of the modification parameters tends to generate modification parameters that modify the audio signal to reduce the difference between the specific loudness and the target specific loudness.

In a playback environment containing a secondary interfering sound signal, such as noise, Calculate the modification parameters, 10, (and its equivalent operation sequences or devices 10 ', 10' 'and 10' '' in each of examples 2-4, respectively), can also take as an optional input signal an indicator of such a secondary interfering audio signal or the secondary interfering signal itself as one of its input signals. Such an optional input signal is shown in FIG. 1 (and FIGS. 2-4) with a dashed plumb line. The secondary interfering signal may be pumped, such as in the example of FIG. 17, described below. Application of the indicator of the interfering signal or the signal itself (it is assumed that the interfering signal is separately accessible for processing) to the sequence of operations or devices 10 Calculate the modification parameters in Fig. 1 (and their equivalent sequences of operations or devices 10 ', 10' 'and 10' '' in each of the examples of FIGS. 2-4, respectively), it is possible to properly configure such a sequence of operations or a device to calculate modification parameters that take into account the interfering signal, as an additional belt below under the heading “Noise Compensation”. In the examples of FIGS. 2-4, the calculation of the specific volume of partial tones suggests that a suitable indicator of the interfering signal is input not only into the corresponding Calculate modification parameters, 10 ', 10' 'or 10' '', but also into the sequence of operations or device 12 “Calculate the approximation of the specific loudness of unmodified audio” and either the flow of operation or device 14 “Calculate the approximation of the target specific loudness” in order to facilitate the calculation of the specific loudness of partial tones by such a function or device ystvom. In the direct-link example of FIG. 1, the specific volume of the partial tones is not calculated explicitly - Calculate the modification parameters, 10, FIG. 1 calculates the appropriate modification parameters to make the specific volume of the partial tones of the modified audio close to the target specific volume. This is further explained below under the heading “Noise Compensation” mentioned above.

As mentioned above, in each of the examples of FIGS. 1-4, the modification parameters M, when applied to the audio signal by the Audio modifier 2, reduce the difference between the specific volume or specific volume of the partial tones of the resulting modified audio and the target specific volume. Specifically, the specific volume of the modified audio signal is well close to or is the same as the target specific volume. Modification parameters M, for example, take the form of time-dependent gain factors applied to frequency bands obtained from a filter bank or to coefficients of a time-dependent filter. Accordingly, in all the examples of Figs. filter).

Here and elsewhere in this document, the use of the same reference number indicates that the device or process may be substantially identical to another or another bearing the same reference number. Link numbers bearing numbers with a dash (for example, “10 '”) indicate that the device or process is similar in design or function, but may be a modification of another or others bearing the same base link number or its variants marked with a dash .

Under certain restrictions, an almost equivalent feedback arrangement can be implemented according to the example with direct coupling of FIG. 1. Figure 2 depicts such an example in which the audio signal is also fed into the sequence of operations or the device 2 Modify the audio signal in the signal path. The sequence of operations or device 2 also receives the modification parameters M from the control path, in which the sequence of operations or device 4 'Generate modification parameters in the feedback arrangement receives a modified sound signal from the output as its input signal. Modify the audio signal, 2. Thus, in the example of FIG. 2, modified audio rather than unmodified audio is fed into the control path. The sequence of operations or device 2 Modify the audio signal and the sequence of operations or device 4 'Generate modification parameters modify the audio signal to reduce the difference between its specific volume and target specific volume. The sequence of operations or device 4 'may include several functions and / or devices: the sequence of operations or device 12 "Calculate the approximation of the specific volume of unmodified audio", the sequence of operations or device 14 "Calculate the approximation of the target specific volume" and the sequence of operations or device 10' “Calculate modification parameters” that calculate the modification parameters.

With the limitation that the function or functions F are reversible, the sequence of operations or device 12 estimates the specific volume of the unmodified audio signal by applying the inverse function F ^-1 to the specific volume or specific volume of the partial tones of the modified audio signal. The device or sequence of operations 12 can calculate the inverse function F ^-1 , as described above. This is schematically shown in FIG. 2 as an input signal, “Select the inverse function (s) F ^-l and function parameters (s)” in the sequence of operations or device 12. “Calculate the approximation of the target specific volume” 14 operates by calculating the specific volume or specific volume of partial tones of the modified audio signal. Such specific volume or specific volume of partial tones is an approximation of the target specific volume. The approximation of the specific volume of the unmodified sound signal and the approximation of the target specific volume are used by Calculate modification parameters, 10 ', to obtain modification parameters M, which, if applied to the audio signal by Modify the audio signal, 2, reduce the difference between the specific volume or specific volume of partial tones modified audio signal and target specific volume. As mentioned above, these modification parameters M, for example, can take the form of time-dependent gain factors applied to the filter bank frequency bands or time-dependent filter coefficients. In practical embodiments, Calculate modification parameters, a 10 ″ feedback loop may introduce a delay between the calculation and application of the modification parameters M.

As mentioned above, in a transmission medium containing a secondary interfering audio signal, such as noise, each of Calculate modification parameters, 10 ', Calculate approximation of specific volume of unmodified audio, 12, and Calculate approximation of target specific volume, 14, can also be taken as optional input signal indicator of such a secondary interfering sound signal or the secondary interfering signal itself as one of its input signals, and the sequence of operations or device 12, and the sequence Each operation or device 14 each can calculate the specific volume of the partial tones of the modified audio signal. Such optional input signals are shown in FIG. 2 using dashed plumb lines.

As mentioned above, examples of mixed direct feedback / feedback implementations of aspects of the invention are possible. Figures 3 and 4 show two examples of such implementations. In the examples of FIGS. 3 and 4, as in the examples of FIGS. 1 and 2, the sound signal is also fed into the sequence of operations or device 2 Modify the sound signal in the signal path, but Generate modification parameters (4 '' in Figs. 3 and 4 ' '' in Fig. 4) in the respective control paths each receive both an unmodified sound signal and a modified sound signal. In both examples, FIGS. 3 and 4, Modify the audio signal, 2, and Generate modification parameters (4 '' and 4 '' ', respectively) modify the audio signal to reduce the difference between its specific volume, which may be implicit, and the target specific volume, which can also be implicit.

In the example of FIG. 3, the flowchart or device 4 'Generate modification parameters may include several functions and / or devices: Calculate the target specific volume, 6, as in the example of FIG. 1, Calculate the approximation of the target specific volume, 14, as in the example 2, and the flowchart or device 10 ″ “Calculate modification parameters”. As in the example of FIG. 1, in the direct-coupling part of this example with mixed direct-feedback / feedback, Calculate the target specific volume, 6, can perform one or more functions “ F ”, each of which may have function parameters. This is schematically shown in FIG. 3 as an input signal, “Select function (s) F and parameter (s) of function (s)” in a flowchart or device 6. In the feedback part of this example with mixed direct feedback / feedback, the modified an audio signal is supplied to Calculate the approximation of the target specific volume, 14, as in the feedback example of FIG. 2. The sequence of operations or device 14 operates in the example of figure 3, as it does in the example of figure 2, calculating the specific volume or specific volume of partial tones of the modified audio signal. Such specific volume or specific volume of partial tones is an approximation of the target specific volume. The target specific volume (from the flow of operations or device 6) and the approximation of the target specific volume (from the flow of operations or device 14) are calculated in Calculate modification parameters, 10 '', to obtain the modification parameters M, which, if applied to the audio signal by Modify the sound signal 2 reduces the difference between the specific volume of the unmodified audio signal and the target specific volume. As mentioned above, these modification parameters M, for example, can take the form of time-dependent gain factors applied to the filter bank frequency bands or time-dependent filter coefficients. In practical embodiments, the feedback loop may introduce a delay between the calculation and application of the modification parameters M. As mentioned above, in a transmission medium containing a secondary interfering audio signal, such as noise, each of Calculate modification parameters, 10``, and Calculate the approximate target specific volume, 14, can also take as an optional input an indicator of such secondary interfering audio signal or the secondary interfering signal itself as one of its input signals, and the sequence of operations or device 14 can calculate the specific volume of partial tones of the modified sound signal Nala. Optional input signals are shown in FIG. 3 using dashed plumb lines.

Calculate the modification parameters, 10 '', can use the device or the error detection function so that the differences between its input signals of the target specific volume and the approximation of the target specific volume adjust the modification parameters in order to reduce the differences between the approximation of the target specific volume and the “real” target specific volume. Such settings reduce the differences between the specific volume of the unmodified audio signal and the target specific volume, which may be implicit. Thus, the modification parameters M can be adapted based on an error between the target specific volume calculated in the direct path from the specific volume of the original audio using function F and the approximation of the target specific volume calculated in the feedback path from the specific volume or specific volume of partial tones modified audio.

In the example of FIG. 4, an alternate example with direct feedback / feedback is shown. This alternative variant differs from the example in FIG. 3 in that the inverse function or functions F ^{−l are} preferable to be calculated in the feedback path than the function or functions F calculated in the direct path. In the example of FIG. 4, the flowchart or device 4 'Generate modification parameters may include several functions and / or devices: Calculate the target specific volume, 8, as in the example with direct connection of FIG. 1, Calculate the approximation of the target specific volume of unmodified audio, 12, as in the feedback example of FIG. 2, and “Calculate modification parameters”, 10 ″. Calculate the specific volume, 8, as in the example with direct connection of Fig. 1, gives as input the signal Calculate modification parameters, 10 ''', the specific volume of the unmodified sound signal. As in the feedback example of FIG. 2, with the limitation that the function or functions F are reversible, the sequence of operations or device 12 estimates the specific volume of the unmodified audio signal by applying the inverse function F ^-1 to the specific volume or specific volume of partial tones of the modified audio signal. The input signal "Select the inverse function (s) and parameter (s) of the inverse function (s)" for Calculating the approximate specific volume of unmodified audio, 12, is shown to indicate that the device or sequence of operations 12 can calculate the inverse function F ^-1 , as described above. This is schematically shown in FIG. 4 as an input signal, “Select the inverse function (s) F ^-l and parameter (s) of the function (s)” to the flowchart or device 12. Thus, the flowchart or device 12 outputs, as another input signal in Calculate modification parameters, 10 ''', approximation of the specific volume of the unmodified sound signal.

As in the examples of FIGS. 1-3, Calculate modification parameters, 10`` ', obtains modification parameters M, which, if applied to the audio signal by Modify the audio signal, 2, reduce the difference between the specific volume of the unmodified audio signal and the target specific volume which is implicit in this example. As mentioned above, the modification parameters M, for example, can take the form of time-dependent gain factors applied to the frequency bands of the filter bank or the coefficients of a time-dependent filter. In practical embodiments, the feedback loop may introduce a delay between the calculation and application of the modification parameters M. As mentioned above, in a transmission medium containing a secondary interfering audio signal, such as noise, each of Calculate modification parameters, 10``, and Calculate the approximate specific volume of unmodified audio, 12, can also take an indicator of such secondary as an optional input signal interfering sound signal or the secondary interfering signal itself as one of its input signals, and the sequence of operations or device 12 can calculate the specific volume of the partial tones of the modified sound signal. Optional input signals are shown in FIG. 4 using dashed plumb lines.

Calculate the modification parameters, 10 ''', can use the device or the error detection function, so that the differences between its input signals of specific volume and the approximation of specific volume form output signals that adjust the parameters of the modification in order to reduce the differences between the approximation of the specific volume and "Real" specific volume. Since the specific loudness approximation is obtained from the specific loudness or specific loudness of partial tones of the modified audio, which can be regarded as the approximation of the target specific loudness, such settings reduce the differences between the specific loudness of the modified audio signal and the target specific loudness, which is inherent in the function or functions F ^-1 . Thus, the modification parameters M can be adapted based on an error between the specific volume calculated in the direct connection path from the original audio and the specific volume approximation calculated using the inverse function or functions F ^-1 in the feedback path from the specific volume or specific volume partial tones of modified audio. Thanks to the feedback path, practical implementations can introduce a delay between the calculation and application of the modification parameters.

Although the modification parameters M in the examples of FIGS. 1-4, when brought into the sequence of operations or device 2 Modify the Sound signal, reduce the difference between the specific volume of the sound signal and the target specific volume, in practical embodiments, the corresponding modification parameters generated in response to one the same sound signal may not be identical to each other.

Although uncritical and inconsequential with respect to aspects of the present invention, the calculation of the specific volume of an audio signal or a modified audio signal can advantageously apply the techniques set forth in the aforementioned international patent application No. PCT / US2004 / 016964, published as WO 2004/111964 A2, in which the calculation carries out selection from a group of two or more functions of the specific volume model, one or a combination of two or more functions of the specific volume model, the choice of which is controlled by the input characteristics indicator sound signal. The specific volume description 104 of FIG. 1 below describes such an arrangement.

In accordance with further aspects of the invention, an unmodified sound signal and either (1) modification parameters, or (2) target specific volume or representation of target specific volume (e.g., scale factors explicitly or implicitly used in calculating the target specific volume) can be stored or transmitted for use, for example, in temporally and / or spatially separated devices or sequences of operations. Modification parameters, target specific loudness, or representation of target specific loudness can be determined in any suitable way, for example, as in one example of a direct-coupled, closed-loop and mixed-direct / closed-coupling arrangement of FIGS. 1-4, as described above . In practice, a direct-coupled arrangement, such as in the example of FIG. 1, is the least complex and fastest because it avoids calculations based on a modified audio signal. An example of transmitting or storing unmodified audio and modification parameters is shown in FIG. 5, while an example of transmitting or storing unmodified audio and target specific volume or representing target specific volume is shown in FIG. 6.

An arrangement, such as in the example of FIG. 5, can be used to temporarily and / or spatially separate the application of the modification parameters to the audio signal from the formation of such modification parameters. An arrangement, such as in the example of FIG. 6, can be used to temporarily and / or spatially separate both, form and apply modification parameters, from generating the target specific volume or presenting it. Both types of arrangements make possible simple low-cost reproduction or reception arrangements that avoid the complexity of generating modification parameters or generating target specific volume. Although the layout of the type of FIG. 5 is simpler than the layout of the type of FIG. 6, the layout of FIG. 6 has the advantage that the information required for storage or transmission can be much smaller, in particular when a representation of the target specific volume, such as one or more scale factors, stored or transmitted. Such a reduction in the storage or transmission of information, in particular, may be useful in audio environments with a low bit rate.

Accordingly, additional aspects of the present invention are the provision of a device or a sequence of operations that (1) receive or reproduce from a device or a sequence of operations of storage or transmission, modification parameters M and apply them to an audio signal that is also received, or (2) which receive or reproduce, from a device or a sequence of operations of storage or transmission, the target specific volume or representation of the target specific volume, form a parameter M modification, applying the target specific volume or its representation to an audio signal that is also received (or an indicator of the audio signal, such as its specific volume that can be obtained from the audio signal), and apply the M modification parameters to the received audio signal. Such devices or process sequences may be characterized as decoding process sequences or decoders; while the devices or sequences of operations required to generate stored or transmitted information may be characterized as encoding sequences or encoders. Such coding sequences or encoders are those parts of the example layouts of FIGS. 1-4 that are used to generate the information required by respective decoding sequences or decoders. Such decoding processors or decoders may be associated with or working with virtually any type of process or device that processes and / or reproduces sound.

In one aspect of the invention, as in the example of FIG. 5, an unmodified sound signal and modification parameters M, for example, generated by a sequence of generation steps or a modification parameter generator, such as Generate modification parameters, 4, in FIGS. 1, 4 'in FIG. .2, 4 ″ in FIG. 3 or 4 ″ ″ in FIG. 4 can be input to any suitable storage or transmission device or function 16 (“Save or Transmit”). In the case of using the direct-link example of FIG. 1 as an encoding or encoder flowchart, Modify the audio signal, 2, would not be required to generate modified audio and could be omitted if there is no need to provide modified audio at the temporal or spatial location of the encoder or coding process. Save or transmit, 16, for example, may include any suitable optical or solid state storage and playback devices or any suitable wired or wireless transmission and reception devices, their selection is not critical to the invention. The reproduced or received modification parameters can then be input to Modify the audio signal, 2, of the type used in the examples of FIGS. 1-4, in order to modify the reproduced or received audio signal, so that its specific volume approaches the target specific volume, or is inherent layout in which the modification parameters were obtained. Modification parameters may be stored or transmitted in any of various ways. For example, they can be stored or transmitted as metadata accompanying an audio signal, they can be sent in separate paths or channels, they can be shorthand encoded into audio, they can be multiplexed, etc. The use of modification parameters to modify the audio signal may be optional and, if not, their use may be selectable, for example, by the user. For example, modification parameters, if applied to an audio signal, could reduce the dynamic range of the audio signal. Whether or not to apply such a reduction in dynamic range could be user selectable.

In yet another aspect of the invention, as in the example of FIG. 6, the unmodified audio signal and the target specific volume or representation of the target specific volume can be supplied to any suitable storage or transmission device or function 16 (“Save or Transmit”). In the case of using a direct-coupling configuration, such as the example of FIG. 1, as a coding or encoder flowchart, neither a flowchart or device of type Calculate modification parameters, 10 nor a flowchart or device of type Modify an audio signal, 2, was required could and could be omitted if there is no need to provide modification parameters or modified audio at the temporal or spatial location of the encoder or op sequence coding radios. As in the case of the example of FIG. 5, Save or Transmit, 16, for example, may include any suitable optical or solid state storage and playback devices, or any suitable wired or wireless transmission and reception devices, their selection is not critical to the invention . The reproduced or accepted target specific volume or representation of the target specific volume can then be input, along with unmodified audio, to Calculate modification parameters, 10, of the type used in the examples of FIG. 3, in order to provide modification parameters M, which can then be input to Modify an audio signal, 2, of the type used in the examples of FIGS. 1-4, in order to modify a reproduced or received signal, so that its specific volume approaches the target specific loudly or inherent in the layout in which the modification parameters were obtained. Although the target specific loudness or its representation can most easily be obtained in the encoding process or an encoder of the approximate type of FIG. 10, the target specific loudness or its representation or approximation to the target specific loudness or its representation can be obtained in the encoding process or the encoder of the exemplary types of FIG. 2 to 4 (approximations are calculated in the sequence of operations or devices 14 of FIGS. 2 and 3 and in the sequence of operations or device 12 of FIG. 4). The target specific volume or its representation may be stored or transmitted in any of various ways. For example, it can be stored or transmitted as metadata accompanying an audio signal, it can be sent in separate paths or channels, it can be shorthand encoded into audio, it can be multiplexed, etc. The use of modification parameters obtained from stored or transmitted target specific volume, or representations for modifying an audio signal may be optional and, if not necessary, their use may be selected, for example, by the user. For example, modification parameters, if applied to an audio signal, could reduce the dynamic range of the audio signal. Whether or not to apply such a reduction in dynamic range could be user selectable.

When implementing the disclosed invention as a digital system, a direct-coupled configuration is most practical, and therefore examples of such configurations are described in detail below, it is understood that the scope of the invention is not so limited.

Throughout this document, terms such as “filter” or “filter comb” are used throughout this application to include essentially any kind of recursive and non-recursive filtering, such as IIR filters or transforms, and “filtered” information is the result of applying such filters. The embodiments described below apply filter banks implemented by transformations.

7 depicts large details of an exemplary embodiment of an aspect of the invention embodied in a direct-coupled arrangement. The audio first passes through a function or device 100 of an analysis filter bank (“Analysis filter bank”) that divides an audio signal into a plurality of frequency bands (hence, FIG. 5 shows multiple output signals from an analysis filter bank, 100, each output signal represents a frequency band, which the output signal carries through various functions or devices up to the synthesis filter bank, which sums the bands into an integrated broadband signal, as further described below). A filter characteristic associated with each frequency band in the Analysis Filter Comb, 100, is intended to simulate a characteristic at a particular location of the basilar membrane in the inner ear. The output of each filter in the Analysis Filter Comb, 100, then passes to a pass filter or pass filter function 101 (“Pass Filter”), which mimics the filtering effect of transmitting audio through the outer and middle ear. If only the audio volume was to be measured, a transmission filter could be applied before the analysis filter bank, but since the output signals of the analysis filter bank are used to synthesize modified audio, it is useful to use a transmission filter after the filter bank. The output signals of the Pass Filter, 101, then pass into a pump function or device 102 (“Pump”), the output signals of which simulate the distribution of energy across the basilar membrane. The values of the pump energy can be smoothed over time by a function or a smoothing device 103 (“Smoothing”). The time constants of the smoothing function are set in accordance with the needs of the desired application. Smoothed pump signals are essentially converted to specific volume in a specific volume function or device 104 (“Specific Volume (SL)”). Specific loudness is presented in units of sleep per unit frequency. A specific loudness component associated with each band passes into a specific loudness modification function or device 105 (“SL Modification”). Modification 105 SL takes as its input a source specific loudness and then outputs the desired or “target” specific loudness, which according to an aspect of the present invention is preferably a function of the original specific loudness (see the following heading entitled “Target specific loudness”). Modification SL, 105 may work independently in each band, or there may be an interdependence between or among the bands (frequency smoothing, as suggested by the cross-connection lines in Fig. 7), depending on the desired result. Taking as its input signals the smoothed components of the pump frequency bands from Pumps, 102, and the target specific volume from Modification SL, 105, the function or device 106 of the gain coefficient (“Gain Solver”) determines the gain to be applied to each band output signal Combs analysis filters, 100, in order to convert the measured specific volume to the target specific volume. The gain factor solver can be implemented in various ways. For example, the Gain Solver may include an iterative process such as in the method disclosed in the aforementioned International Patent Application No. PCT / US2004 / 016964 published as WO 2004/111964 A2, or as an alternative to a lookup table. Although the band coefficients generated by the Gain Solver, 106, can be further time-smoothed by an optional smoothing function or device 107 (“Smoothing”) in order to minimize perception artifacts, it is preferable that temporal smoothing is applied elsewhere in the whole a process or device as described elsewhere. In conclusion, the amplification factors are applied to the respective bands of the Analysis filter comb, 100 through the corresponding multiplicative combining function or combiner 108, and the processed or “modified” audio is synthesized from the modified gain bands in the synthesis filter comb function or device 110 (“Synthesis filter comb”) . In addition, the output from the analysis filter bank may be delayed by a delay function or device 109 (“Delay”) until the amplification factors are applied in order to compensate for any delay associated with the calculation of the amplification factors. Alternatively, instead of calculating gain factors to use when applying gain bandwidth modifications, gain resolvers, 106, can calculate filter coefficients that control a time-dependent filter, such as a multi-tap FIR filter or a multi-pole IIR filter. For simplicity of description, aspects of the invention are mainly described as using gain factors applied to frequency bands, it is understood that filter coefficients and time-dependent filters can also be applied in practical embodiments.

In specific embodiments, audio processing may be performed in the digital domain. Accordingly, the audio input signal is indicated by a discrete time sequence x [ n ], which was sampled from the audio source at a certain sampling frequency f _s . It is assumed that the sequence x [ n ] was appropriately scaled so that the rms power x [ n ] in decibels given by

equal to the sound pressure level in dB at which the audio is heard by a human listener. In addition, an audio signal is assumed to be monophonic for ease of description.

Analysis Filter Comb, 100, Pass Filter, 101, Pumping, 102, Specific Volume, 104, Specific Volume Modification, 105, Gain Solver, 106, and Synthesis Filter Comb, 110, can be described in more detail as follows.

Filter Comb, 100

The audio input signal is fed into the filter bank or function 100 of the analysis filter bank (“Analysis filter bank”). Each filter in the Analysis Filter Comb, 100, is designed to simulate the frequency response at a specific location along the basilar membrane in the inner ear. Filter bank 100 may include a set of linear filters whose bandwidth and spacing are constant on an equivalent rectangular bandwidth (ERB) scale as defined by Moore, Glasberg and Baer (BCJ Moore, B. Glasberg, T. Baer, “A Model for the Prediction of Thresholds, Loudness, and Partial Loudness ”above).

Although the ERB frequency scale is closer to human perception and shows improved performance when creating measurements of objective volume that correspond to the results of subjective volume, the Bark frequency scale can be used with reduced quality of operation.

For the center frequency f in hertz, the width of one ERB band in hertz can be approximately expressed as:

This dependence determines the curvilinear frequency scale from the condition that any point on the curvilinear scale corresponding to ERB in units of the curvilinear scale is equal to unity. The function for converting from linear frequency in Hertz to this ERB frequency scale is obtained by integrating the analogue of equation 1:

It is also useful to express the conversion from the ERB scale back to the linear frequency scale by solving equation 2a with respect to f :

where e takes place in units of the ERB scale. Figure 9 shows the relationship between the ERB scale and frequency in Hertz.

The analysis filter bank, 100, may include B auditory filters, referred to as bands, at center frequencies f _c [1] ... f _c [ B ] uniformly spaced along the ERB scale. More accurately,

where ∆ is the required spacing of the ERB Combs of the analysis filters, 100 and where f _min and f _max are the required minimum and maximum center frequencies, respectively. You can choose ∆ = 1, and, taking into account the frequency range over which the human ear is sensitive, you can set f _min = 50 Hz and f _max = 20,000 Hz . With such parameters, for example, applying equations 3a-c gives B = 40 auditory filters.

The frequency response of each auditory filter may be characterized by a spherical exponential function, as proposed by Moore and Glazberg. More precisely, the amplitude characteristic of the filter with a central frequency f _c [ b ] can be calculated as:

Where

The amplitude characteristics of such B auditory filters, which are approximately equal to the definition of critical bands on the ERB scale, are shown in FIG. 10.

Filtering operations The comb filter analysis, 100, can be satisfactorily approximately expressed using a finite-length discrete Fourier transform, usually referred to as a short-term discrete Fourier transform (STDFT), since an implementation that performs filters at the audio sampling frequency, indicated by the link as a full-frequency implementation, It is supposed to give a greater temporal resolution than is necessary for accurate volume measurements. By using STDFT instead of full-frequency implementation, improved efficiency and reduced computational complexity can be achieved.

The STDFT of the audio input signal x [ n ] is defined as:

where k is the frequency index, t is the time interval index, N is the size of the DFT (discrete Fourier transform, DFT), T is the size of the jump, and w [ n ] is the length of the window N , normalized so that

Note that the variable t in equation 5a is a discrete index representing the time interval STDFT, as opposed to measuring time in seconds. Each increment in t represents a jump of T samples along the signal x [ n ] . Subsequent references to the index t imply this definition. Despite the fact that different parameter settings and window shapes can be used depending on the implementation details, for f _s = 44100 Hz , choosing N = 2048, T = 1024, and forcing w [ n ] to be a Hanning window provides a sufficient balance of time and frequency resolution . The STDFT described above can be more efficient using Fast Fourier Transform (FFT).

Instead of STDFT, a modified discrete cosine transform (MDCT) can be used to implement an analysis filter bank. The MDCT conversion commonly used in audio-related encoders, such as Dolby's AC-3 systems. If the disclosed system is implemented using such perceptually encoded audio, the disclosed volume measurement and modification can be more effectively implemented by processing the existing MDCT coefficients of the encoded audio, thereby eliminating the need to convert the analysis filter bank. The MDCT of the audio input signal x [ n ] is defined by:

Usually, the jump size T is chosen to be exactly half the length N of the transform, so that perfect reconstruction of the signal x [ n ] is possible.

Pass Filter, 101

The output signals of the Analysis Filter Combs 100, are fed into a transmission filter or a transmission filter function 101 (“Transmission Filter”), which filters each band of the filter comb according to audio transmission through the outer and middle ear. Fig. 8 depicts one of the suitable amplitude-frequency characteristics of a transmission filter, P ( f ), over a bandwidth of audible frequencies. The characteristic is a unit below 1 kHz and above 1 kHz, followed by an inversion of the auditory threshold, which is specified in the ISO226 standard, with a threshold value normalized to equal unity at 1 kHz.

Pumping, 102

In order to calculate the volume of the input sound signal, an indicator of short-term energy of sound signals in each filter is required. The filter banks of the analysis filters, 100, after applying the Pass filter, 101. This time and frequency-dependent indicator is indicated by reference as pumping. The output signal of the short-term energy of each filter in the Analysis Filter Comb, 100, can be approximately expressed in the Pump Functions, 102, by multiplying the characteristics of the filters in the frequency domain with the power spectrum of the input signal:

where b is the band number, t is the frame number, and H _b [ k ] and P [ k ] are the frequency characteristics of the auditory filter and transmission filter, respectively sampled at a frequency corresponding to the index k of the STDFT or MDCT receiver. It should be noted that varieties for the amplitude characteristics of auditory filters other than those specified in equations 4a-c can be used in equation 7 to achieve similar results. For example, the aforementioned international application No. PCT / US2004 / 016964, published as WO 2004/111964 A2, describes two alternatives: an auditory filter characterized by a ^12th order IIR transform function, and an approximation of an inexpensive “sharp cut” bandpass filter.

So, the output signal of the Pump Function, 102, is a representation in the frequency domain of the energy E in the respective bands b ERB over a period of t time.

Time Averaging (“Smoothing”), 103

For some applications of the disclosed invention, as described below, it may be desirable to smooth the pump E [ b , t ] before converting it to a specific loudness. For example, smoothing can be performed recursively in smoothing function 103 according to the equation:

where the time constants λ _b in each band b are selected in accordance with the desired application. In most cases, time constants can advantageously be chosen to be proportional to the integration time of human perception of loudness within band b. Watson and Handel performed experiments demonstrating that this integration time was within the range of 150-175 ms at low frequencies (125-200 Hz) and 40-60 ms at high frequencies (Charles S. Watson and Roy W. Gengel, “ Signal Duration and Signal Frequency in Relation to Auditory Sensitivity Journal of the Acoustical Society of America, Vol. 46, No. 4 (Part 2), 1969, pp. 989-997 (Charlie S., Watson and Roy W. Handel, “Signal duration and signal frequency relative to auditory sensitivity” ( Journal of the Acoustic Society of America , Volume 46, No. 4 (Part 2), 1969, pp. 989-997)).

Specific Volume, 104

In the converter or specific volume conversion function 104 (“Specific Volume”), each pump band is converted to a specific volume component, which is measured in sones on the ERB.

[b, t] может преобразовываться в эквивалентный уровень накачки на 1 кГц, как задано кривыми равной громкости по ISO 226 (фиг.11), нормализованным пропускающим фильтром P(z) фиг.12):Initially, when calculating the specific volume, the pump level in each band

[ b , t ] can be converted to an equivalent pump level at 1 kHz, as defined by equal-volume curves in accordance with ISO 226 (FIG. 11), normalized by a transmission filter P (z) of FIG. 12):

where T _1kHz ( E , f ) is a function that forms a level at 1 kHz, which is equal in volume to level E at a frequency f . In practice, T _1kHz ( E , f ) is implemented as interpolation of the reference table of equal volume curves normalized by a transmission filter. Converting equivalent levels to 1 kHz simplifies the subsequent calculation of specific volume.

Then the specific volume in each band can be calculated as:

where N _NB [ b , t ] and N _WB [ b , t ] are the specific volume values based on the narrow-band or wide-band signal transmission model, respectively. The value α [ b , t ] - is the interpolation coefficient lying between 0 and 1, which is calculated from the sound signal. Mentioned international application No. PCT / US2004 / 016964, published as WO 2004/111964 A2, describes a technology for calculating α [ b , t ] from spectral pump irregularities. It also describes in more detail the "narrow-band" and "broadband" signal transmission models.

The values of N _NB [ b , t ] and N _WB [ b , t ] of narrowband and broadband specific volume can be estimated from the converted pump using exponential functions:

where TQ _1kHz is the pump level at a threshold value in silence for a tone of 1 kHz. According to the equal volume curves (Figs. 11 and 12), TQ _lkHz is 4.2 dB. We note that both of these specific volume functions are equal to zero when the pump is equal to the threshold value in silence. For pumps larger than the threshold value in silence, both functions monotonically increase with a power dependence in accordance with Stevens law on the feeling of intensity. The exponent for the narrowband function is selected to be greater than that of the broadband function, causing the narrowband function to increase faster than the broadband function. A separate set of exponents β and gain G for narrowband and wideband cases is selected to match experimental data on volume growth for tones and noise.

Moore and Glasberg put forward as an assumption that the specific volume should be equal to some small value instead of zero when the pump is at the threshold of hearing. The specific volume should then decrease monotonously to zero as the pump decreases to zero. The rationale is that the auditory threshold is a probabilistic threshold (the point at which a tone is detected 50% of the time) and that the number of tones, each at a threshold value presented together, can be summed into a sound that we hear more than any of the individual tones. In the disclosed application, the replenishment of specific volume functions with this property has the additional advantage of inducing the gain factor solver, discussed below, to behave more appropriately when the pump is near a threshold value. If the specific volume is determined to be zero when the pump is at or below a threshold value, then there is no unique solution for the gain factor solver for pumping at or below a threshold value. If, on the other hand, the specific loudness is determined to increase monotonically for all pump values greater than or equal to zero, as proposed by Moore and Glazberg, then a unique solution exists. Scaling a volume larger than one will always result in a gain greater than one, and vice versa. The specific volume functions in equations 11a and 11b can be modified to have the desired property, according to:

_1kHz[b, t] = λTQ _1kHz.where the constant λ is greater than one, the exponent η is less than one, and the constants K and C are chosen so that the specific volume function and its first derivative are continuous at

_1kHz [ b , t ] = λTQ _1kHz .

From the specific volume, the total or “full” volume L [ t ] is set by the sum of the specific volume for all bands 6:

Specific Volume Modification, 105

[b, t], может рассчитываться по удельной громкости из SL 104 (фиг.7) различными способами, зависящими от требуемого применения всего устройства или последовательности операций. Как более подробно описано ниже, целевая удельная громкость может рассчитываться с использованием масштабного коэффициента α, например, в случае регулировки уровня громкости. Смотрите уравнение 16, приведенное ниже, и его ассоциативно связанное описание. В случае автоматической регулировки усиления (АРУ) и регулировки динамического диапазона (DRC) целевая удельная громкость может рассчитываться с использованием отношения требуемой выходной громкости ко входной громкости. Смотрите уравнения 17 и 18, приведенные ниже, и их ассоциативно связанные описания. В случае динамической коррекции целевая удельная громкость может рассчитываться с использованием зависимости, изложенной в уравнении 23, и ее ассоциативно связанном описании.In the specific volume modification function 105 (“Specific Volume Modification”), the target specific volume indicated by reference as

[ b , t ], can be calculated from the specific volume from SL 104 (Fig. 7) in various ways, depending on the desired application of the entire device or sequence of operations. As described in more detail below, the target specific volume can be calculated using a scale factor α , for example, in the case of adjusting the volume level. See equation 16 below and its associated description. In the case of automatic gain control (AGC) and dynamic range control (DRC), the target specific volume can be calculated using the ratio of the desired output volume to the input volume. See equations 17 and 18 below and their associated descriptions. In the case of dynamic correction, the target specific loudness can be calculated using the dependence set forth in equation 23 and its associated description.

Gain Solver, 106

[b, t] и целевую удельную громкость

[b, t] и формирует коэффициенты G[b, t] усиления, используемые впоследствии для модификации аудио. С допущением, что функция Ψ{·} представляет нелинейное преобразование из накачки в удельную громкость, из условия чтобыIn this example, for each band b and each time interval t , the Gain Solver, 106, receives a smooth pump as its input signal

[ b , t ] and target specific volume

[ b , t ] and forms the gain factors G [ b , t ], which are subsequently used to modify the audio. Assuming that the function Ψ {·} represents a nonlinear transformation from pump to specific loudness, so that

The gain solver finds G [ b , t ], so that

Gain solvers, 106, determine frequency and time-dependent gain factors, which, when applied to the initial pump, result in a volume that is ideally equal to the desired target specific volume. In practice, Gain Solvers determine the frequency and time-dependent gain factors that, when applied to a variation in the frequency domain of the audio signal, result in a modification of the audio signal in order to reduce the difference between its specific loudness and target specific loudness. Suspiciously, the modification is such that the modified sound signal has a specific volume, which is a good approximation of the target specific volume. The solution to equation 14a can be implemented in a variety of ways. For example, if there is a mathematical expression reflecting a number of solutions for the inversion of the specific volume represented by Ψ ^-1 {·}, then the gain can be calculated directly by rearranging equation 14a:

Alternatively, if there is no reflective series of solutions for the mathematical expression for Ψ ^-1 {·}, an iterative approach can be used in which, during each iteration, equation 14a is estimated using the current gain estimate. The resulting specific volume is compared with the desired target, and the gain is updated based on the error. If the gains are updated properly, they will converge to the desired solution. Another method involves pre-calculating the function { ^-1 {·} for the range of pump values in each strip to create a look-up table. An approximation of the inverse function { ^-1 {·} is obtained from this look-up table, and the amplification factors, in this case, can be calculated from equation 14b. As mentioned earlier, the target specific volume can be represented by scaling the specific volume;

Substituting equation 13 into 14c and then 14c into 14b gives an alternative expression for the gains:

[b, t] и масштабирования Ξ [b, t] удельной громкости. Поэтому, коэффициенты могут вычисляться посредством оценки по 14d или эквивалентной справочной таблицы без явного вычисления когда-либо удельной громкости или целевой удельной громкости в качестве промежуточных значений. Однако, эти значения вычисляются неявно благодаря использованию уравнения 14d. Могут быть изобретены другие эквивалентные способы для вычисления параметров модификации через явное или неявное вычисление удельной громкости и целевой удельной громкости, и это изобретение подразумевается покрывающим все такие способы.We see that the gains can be perfectly expressed as a pump function

[ b , t ] and scaling Ξ [ b , t ] specific volume. Therefore, the coefficients can be calculated by estimating from 14d or an equivalent lookup table without explicitly calculating the ever specific loudness or target specific loudness as intermediate values. However, these values are calculated implicitly by using equation 14d. Other equivalent methods can be invented for calculating the modification parameters by explicitly or implicitly calculating the specific volume and the target specific volume, and this invention is intended to cover all such methods.

Synthesis Filter Comb, 110

As described above, the Analysis Filter Comb, 100, can be efficiently implemented using either the Short-Term Discrete Fourier Transform (STDFT) or the Modified Discrete Cosine Transform, and STDFT or MDCT can be used in a similar way to implement the Synthesis Filter Comb, 110. More precisely, with the assumption that X [ k , t ] represents the STDFT or MDCT of the input audio, as previously defined, the STDFT or MDCT of the processed (modified) audio in the Synthesis Filter Comb, 110, can be calculated as

where S _b [ k ] is the synthesis filter characteristic associated with band b , and d is the delay associated with delay unit 109 in FIG. 7. The shape of the synthesis filters S _b [ k ] can be selected to be the same as the filters used in the analysis filter bank, H _b [ k ], or they can be modified to provide flawless reconstruction in the absence of any modification by amplification factors (that is, when G [ b , t ] = 1). The final processed audio, in this case, can be formed by the inverse Fourier transform or the modified cosine transform X [ k , t ] and synthesis with the addition of overlapping, as is familiar to a person skilled in the art.

Target specific volume

[b, t]. Хотя изобретение не ограничено никакими конкретными функцией или обратной функцией для расчета целевой удельной громкости, несколько таких функций и подходящих применений для них будут описаны далее.The behavior of arrangements embodying aspects of the invention, such as the examples of FIGS. 1-7, is mainly dictated by the manner in which the target specific volume is calculated

[ b , t ]. Although the invention is not limited to any specific function or inverse function for calculating the target specific volume, several such functions and suitable applications for them will be described later.

Time-independent and frequency-independent function suitable for volume control

A standard volume control adjusts the volume of the audio signal by applying wideband gain to audio. Typically, the gain is tied to a knob or slider that the user adjusts until the audio volume is at the desired level. An aspect of the present invention provides a more psychoacoustic compatible method for implementing such a regulator. According to this aspect of the invention, it is preferable to have a broadband gain associated with a volume control that results in a change in the gain by the same amount across all frequency bands, which can cause a change in the perceived spectrum, the specific volume scaling factor is instead associated with adjusting the control volume level so that the gain on each of the multiple frequency bands changes by an amount that I used to read a model of the human ear, so that in accordance with the ideal there is no change in the perceived spectrum. In the context of this aspect of the invention and its approximate application, “permanent” or “time independent” is intended to include changes to the scaling factor of the volume control, from time to time, for example, by the user. Such “time independence” is sometimes referred to as “almost time-independent”, “quasistationary”, “piecewise time-independent”, “piecewise stationary”, “step-wise time-independent” and “step-wise stationary”. Given such a scale factor, α , the target specific volume can be calculated as the measured specific volume multiplied by α .

Since the full volume L [ t ] is the sum of the specific volume N [ b , t ] over all b bands, the above modification also scales the full volume with the coefficient α , but it does this in a way that preserves the same perceived spectrum at a particular point in time for setting changes volume control. In other words, at any given time, a change in the volume control setting results in a change in the perceived volume, but no change in the perceived spectrum of the modified audio compared to the perceived spectrum of unmodified audio. Fig. 13a depicts the resulting multiband gains G [ b , t ] over all bands “ b ” at a particular point in time “ t ” when α = 0.25 for an audio signal consisting of female speech. For comparison, the broadband gain factor required to scale the original full volume with 0.25 (horizontal line) is also plotted, as in the standard volume control. The multiband coefficient G [ b , t ] increases in the low and high frequency bands compared to the mid-frequency bands. This is compatible with equal volume curves indicating that the human ear is less sensitive at low and high frequencies.

Fig.13b depicts the specific volume for the original audio signal modified with a wideband gain of the signal, which was modified in accordance with the volume control of the prior art, and modified with a multi-band gain of the signal, which was modified in accordance with this aspect of the invention. The specific loudness of a signal modified with a multi-band gain is that of the original signal, scaled by 0.25. The specific volume of a signal modified by a broadband gain has changed its spectral shape with respect to that of the original unmodified signal. In this case, the specific volume, in a relative sense, loses volume at both low and high frequencies. This is perceived as dimming the audio as its volume level decreases, a problem that does not occur with a multi-band modified signal whose volume is controlled by gain factors obtained in the perception loudness region.

[b, t], выбор постоянных λ _b времени в уравнении 8 предписывает, насколько быстро коэффициенты усиления могут меняться во времени в каждой полосе b. Как упомянуто ранее, эти постоянные времени могут выбираться пропорциональными времени интегрирования человеческого восприятия громкости в пределах полосы b и, таким образом, давать надлежащее изменение G[b, t] со временем. Должно быть отмечено, что, если временные ограничения выбираются ненадлежащим образом (слишком быстрыми либо слишком медленными), то неприятные по восприятию артефакты могут привноситься в обработанное аудио.Along with the distortion of the perceived spectral balance associated with the traditional volume control, there is a second problem. The property of loudness perception, which is reflected in the loudness model reflected in equations 11a-11d, is that the volume of the signal at any frequency decreases faster as the signal level approaches the threshold of audibility. As a result, the electrical attenuation required to transmit the same volume attenuation to a quieter signal is less than that required for a louder signal. The traditional volume control gives constant attenuation regardless of signal level, and therefore quiet signals become “too quiet” with respect to louder signals as the volume level decreases. In many cases, this results in loss of detail in the audio. Consider a castanet recording in a reverb room. In such a recording, the main “hit” castanet is too loud compared to the reverberating echoes, but these are reverberating echoes that convey the size of the room. As the volume level is reduced by the traditional volume control, the reverberating echoes become quieter with respect to the main beat and eventually disappear below the audibility threshold, leaving a “dry” sound castanet. The volume-based volume control prevents the quieter parts of the recordings from disappearing by raising the quieter reverberating parts of the recording with respect to the louder main beat, so that the relative volume between these parts remains constant. In order to achieve this effect, the multi-band gains G [ b , t ] must change in time at a rate that is commensurate with the human temporal resolution of loudness perception. Since the multiband gains G [ b , t ] are calculated as a function of the smoothed pump

[ b , t ], the choice of time constants λ _b in equation 8 prescribes how quickly the gains can change in time in each band b. As mentioned earlier, these time constants can be selected proportional to the integration time of human perception of loudness within the band b and, thus, give a proper change in G [ b , t ] over time. It should be noted that if time constraints are chosen improperly (too fast or too slow), then artifacts that are unpleasant in perception can be introduced into the processed audio.

Time-independent and frequency-dependent function suitable for continuous amplitude correction

In some applications, someone may wish to apply a constant amplitude correction of perception to audio, in which case the target specific loudness can be calculated by applying a time-independent but frequency-dependent scale factor Θ [ b ], as a function of

[b, t] - целевая удельная громкость, N[b, t] - удельная громкость звукового сигнала, b - показатель частоты, а t - показатель времени. В этом случае масштабирование может изменяться от полосы к полосе. Такое применение, например, может быть полезным для подчеркивания части спектра, доминируемой речевыми частотами для того, чтобы повышать разборчивость.wherein

[ b , t ] is the target specific volume, N [ b , t ] is the specific volume of the audio signal, b is a frequency indicator, and t is a time indicator. In this case, the scaling can vary from strip to strip. Such an application, for example, may be useful in emphasizing the portion of the spectrum dominated by speech frequencies in order to increase intelligibility.

Frequency-independent and time-dependent function for automatic gain control and dynamic range adjustment

Technologies for automatic gain control and dynamic range control (AGC and DRC) are widely known in the field of sound processing. In an abstract sense, both technologies measure the level of the audio signal in some way, and then modify the signal with a gain by a value that is a function of the measured level. For the AGC case, the signal is modified by the gain, so that its measured level is closer to the control level selected by the user. Using the DRC, the signal is modified by the gain, so that the range of the measured signal level is converted to some desired range. For example, someone might want to make the quieter parts of the audio louder and the louder parts more quiet. Such a system is described by Robinson and Kenneth Gundry (“Dynamic Range Control via Metadata”, 107 ^th Convention of the AES, Preprint 5028, September 24-27, 1999, New York (Charles Robinson and Kenneth Gundry, “Dynamic range by metadata ", the AES 107 ^th Convention, preprint 5028, 24-27 September 1999, New York)). Conventional AGC and DRC implementations typically use simple audio level measurement, such as a smoothed peak or rms amplitude, to produce a gain modification. Such simple measurements can to some extent correlate with the perceived audio volume, but aspects of the present invention provide more perceptible AGCs and DRCs through gain modifications with volume measurements based on the psychoacoustic model. In addition, many conventional AGC and DRC systems use gain modification using wideband gain, thereby entailing the aforementioned tonal (spectral) distortions in the processed audio. Aspects of the present invention, on the other hand, use a multi-band gain to shape the specific volume in some way that reduces or minimizes such distortion.

Both applications, AGC and DRC, applying aspects of the present invention, are characterized by a function that converts or maps the input broadband volume L _i [ t ] to the desired output broadband volume L _o [ t ], where the volume is measured in units of sensory volume, such as sleep. The input broadband volume L _i [ t ] is a function of the specific volume N [ b , t ] of the input audio signal . Although it may be the same as the full volume of the input audio signal, it may be a time-smoothed version of the full volume of the audio signal.

Figa and 14b depict examples of display functions typical of the AGC and DRC, respectively. Under the condition of such a display in which L _o [ t ] is a function of L _i [ t ], the target specific volume can be calculated as

[b, t]. Для системы АРУ входная широкополосная громкость L _i[t] обычно должна быть показателем долговременной полной громкости аудио. Это может достигаться сглаживанием полной громкости L[t] по времени для формирования L _i[t].The original specific volume N [ b , t ] of the audio signal is simply scaled by the ratio of the desired output broadband volume to the input broadband volume to give an output specific volume

[ b , t ]. For an AGC system, the input broadband volume L _i [ t ] should usually be an indicator of the long-term full audio volume. This can be achieved by smoothing the full volume L [ t ] in time to form L _i [ t ].

Compared to the AGC, the DRC system responds to shorter-term changes in the signal volume, and therefore L _i [ t ] can simply be made equal to L [ t ]. As a result, the specific volume scaling specified by L _o [ t ] / L _i [ t ] can quickly fluctuate, leading to undesirable artifacts in the perceived audio. One typical artifact is the audible modulation of part of the frequency spectrum by some other relatively unrelated part of the spectrum. For example, a selection of classical music might contain high frequencies dominated by a long stringed note, while low frequencies contain loudly rumbling timpani. Whenever the timpani strike, the overall volume L _i [ t ] rises, and the DRC system applies the attenuation to the overall specific volume. Strings, in this case, are heard “empty” and filled, depending on the timpani. Such cross-pulsation in the spectrum is also a problem in a traditional system with wideband DRC, and a typical solution entails the use of DRC independently with respect to different frequency bands. The system disclosed here is inherently multi-band due to a filter bank and specific volume calculation, which apply the perception loudness model, and therefore, modifying the DRC system to operate in a multi-band manner in accordance with aspects of the present invention is relatively simple and will be described later.

Time-dependent and frequency-dependent function suitable for dynamic range adjustment

The DRC system can be expanded to operate in a multi-band or frequency-dependent manner by providing input and output volume capabilities to independently vary with band b . These multiband loudness values are referred to as L _i [ b , t ] and L _o [ b , t ], in which case the target specific loudness can be set according to

where L _o [ b , t ] was calculated from or mapped from L _i [ b , t ], as illustrated in FIG. 14b, but independently for each band b . Multiband input loudness L _i [b, t] is a function of the specific loudness N [b, t] of the input audio signal. Although it may be the same as the specific volume of the input audio signal, it may be time-smoothed and / or frequency-smoothed version of the specific volume of the audio signal.

The simplest way to calculate L _i [ b , t ] is to set it equal to the specific volume N [ b , t ] . In this case, the DRC is rather performed independently in each band in the filter bank of the auditory perception model than in accordance with the same input to output volume ratio for all bands, as just described above under the heading “Frequency-independent and time-dependent function suitable for automatic gain control and dynamic range adjustment. " In a practical embodiment employing 40 bands, spacing of these bands along the frequency axis is relatively frequent in order to give an accurate measure of volume. However, applying the DRC scale factor independently to each band can cause the processed audio to sound “fragmented”. To avoid this problem, one may prefer to calculate L _i [ b , t ] by smoothing the specific volume N [ b , t ] in the bands, so that the applied DRC from one band to the next does not change so radically. This can be achieved by determining the filter Q ( b ) to smooth the bands, and then smooth the specific volume for all bands c according to the standard convolution sum:

moreover, N [ c , t ] is the specific volume of the sound signal, and Q ( bc ) is the band-shifted characteristic of the smoothing filter. Fig. 15 shows one example of such a smoothing filter band.

If the DRC function, which calculates L _i [ b , t ] as a function of L _o [ b , t ], is fixed for each band b , then the type of change introduced into each band of specific loudness N [ b , t ] will be vary depending on the spectrum of the audio being processed, even if the overall volume of the signal remains the same. For example, an audio signal with loud bass and a quiet soprano may contain bass cut and soprano increased. A signal with quiet bass and a loud soprano can cause the opposite to happen. The net effect is a change in the timbre or perceived spectrum of the audio, and this may be desirable in certain applications.

[b, t] представляет усредненный по времени вариант многополосной входной громкости L _i[b, t]. Многополосная выходная громкость в таком случае может рассчитываться в качествеHowever, someone may wish to perform multi-band DRC without modifying the normal perceived audio spectrum. Someone might wish that the average modification in each strip was approximately the same, along with the previous provision of the possibility of short-term modifications of the modifications for independent work between or among the bands. The desired effect can be achieved by forcing the average DRC behavior in each band to be the same as that of some average reference behavior. You can select this reference behavior as the desired DRC for the broadband input volume L _i [ t ]. Let the function L _o [ t ] = DRC { L _i [t]} represent the desired DRC mapping for broadband volume. In this case, let

[ b , t ] represents a time-averaged version of the multiband input volume L _i [ b , t ]. In this case, the multiband output volume can be calculated as

Note that the multiband input volume is first scaled to be in the same mid-range as the broadband input volume. Then, the DRC function for broadband volume is applied. Finally, the result is scaled back to the middle range of the multiband volume. With this multi-band DRC formulation, reduced spectral ripple is maintained while preserving the average perceived audio spectrum.

Frequency-dependent and time-dependent function suitable for dynamic correction

[b, t]. Как при многополосной DRC, кто-то может не пожелать, чтобы модификация DEQ радикально менялась от одной полосы к следующей, а потому функция сглаживания полос может применяться для формирования сглаженного по полосам спектра

[b, t]:Another application of aspects of the present invention is the deliberate transformation of a time-dependent perceived audio spectrum into a target time-independent spectrum, while still maintaining the original dynamic range of the audio. This processing may be referred to as dynamic correction (DEQ). In traditional static correction, simple constant filtering is applied to audio in order to change its spectrum. For example, you can apply a constant rise in bass and soprano. Such processing does not take into account the current spectrum of audio, and therefore may not be suitable for some signals, that is, signals that already contain a relatively large amount of bass or soprano. With DEQ, the spectrum of the signal is measured, and the signal is then dynamically modified in order to transform the measured spectrum into a substantially static desired shape. With regard to aspects of the present invention, such a desired shape is defined in stripes in the filter bank and is referred to as EQ [ b ]. In a practical embodiment, the measured spectrum should represent the average spectral shape of the audio, which can be formed by smoothing the specific volume N [ b , t ] over time. Smoothed specific volume may indicate as

[ b , t ]. As with multi-band DRC, someone might not want the DEQ modification to change radically from one band to the next, and therefore the band smoothing function can be used to form a band-smoothed spectrum

[ b , t ]:

[b, t]. Эту нормализованную спектральную форму могут указывать ссылкой как

_EQ[b, t]:In order to maintain the original dynamic range of the audio, the required spectrumEq[b] Should be normalized to have the same overall loudness as the measured spectral shape given by

[b,t]. This normalized spectral shape may be referred to as

_Eq[b,t]:

In conclusion, the target specific volume is calculated as

where β is a user-specified parameter ranging from zero to one, indicating the degree of DEQ to be applied. Looking at equation 23, someone will notice that when β = 0, the initial specific volume is unmodified, and when β = 1, the specific volume is scaled by the ratio of the desired spectral form to the measured spectral form.

[b, t], которая измерена для некоторой части аудио, чей спектральный баланс находит приятным пользователь. В практическом варианте осуществления, например, как показанный на фиг.16, пользователь может снабжаться кнопкой или другим подходящим исполнительным механизмом 507, который, когда приведен в действие, вызывает фиксацию текущего показателя спектральной формы

[b, t] аудио, а затем сохраняет этот показатель в качестве предустановки (в Фиксации и хранении предустановки целевой удельной громкости, 506), которая позже может загружаться в EQ[b], когда задействована DEQ (как посредством выбора 508 предустановки). Фиг.16 - упрощенный вариант фиг.7, в котором только одна линия показана для представления многочисленных полос из Гребенки фильтров анализа, 100, в Гребенку фильтров синтеза, 110. Пример фиг.17 также предусматривает Модификацию удельной громкости (SL) динамической EQ, 505, которая обеспечивает модификацию в отношении удельной громкости, измеренной функцией или устройством 104, в соответствии с динамической коррекцией, как поясненная выше.One of the convenient ways to form the required spectral shape EQ [ b ] is designed so that the user sets it equal

[ b , t ], which is measured for some part of the audio whose spectral balance is pleasant to the user. In a practical embodiment, for example, as shown in FIG. 16, the user may be provided with a button or other suitable actuator 507, which, when activated, causes the current spectral shape to be locked.

[ b , t ] audio, and then saves this metric as a preset (in Fixing and storing the target specific volume preset, 506), which can later be loaded into EQ [ b ] when DEQ is enabled (as by selecting preset 508). FIG. 16 is a simplified version of FIG. 7, in which only one line is shown to represent multiple bands from the Analysis Filter Comb, 100, to the Synthesis Filter Comb, 110. The example of FIG. 17 also provides for Specific Volume (SL) Modification of the dynamic EQ, 505 , which provides a modification with respect to the specific volume measured by the function or device 104, in accordance with the dynamic correction, as explained above.

Combined processing

Someone may wish to combine all the processing described above, including volume control (VC), AGC, DRC and DEQ into a single system. Since each of these sequences of operations can be represented as a specific volume scaling, all of them are easily combined, as follows:

where Ξ _* [ b , t ] represents the scale factors associated with the sequence of operations “*”. A single set of gain factors G [ b , t ] can then be calculated for the target specific volume, which represents the combined processing.

In some cases, the scale factors of one or a combination of volume modification sequences can fluctuate too quickly over time and create artifacts in the resulting processed audio. Therefore, it may be desirable to smooth out some subsets of these scaling factors. In general, the scale factors from VC and DEQ vary uniformly over time, but smoothing the combination of AGC and DRC scale factors may be required. Let a combination of these scale factors be represented by

The main idea behind anti-aliasing is that combined scale factors should respond quickly when the specific volume increases, and that scale factors should smooth out more when the specific volume decreases. This idea is in line with the well-known practice of using rapid onset and slow release in the design of sound compressors. The proper time constants for smoothing the scale factors can be calculated by time-smoothing of the band-specific version of the specific volume. First of all, the specific volume variant smoothed over the bands is calculated:

wherein N [ c , t ] is the specific volume of the sound signal, and Q ( bc ) is the band-shifted characteristic of the smoothing filter, as in equation 19 above.

A time-smoothed version of this band-smoothed specific volume is then calculated as

where the band-dependent smoothing coefficient λ [ b , t ] is set according to

The smoothed combined scale factors are then calculated as

where λ _M [ b , t ] is the band-smoothed version of λ [ b , t ]:

Smoothing by bands of smoothing coefficients prevents time-smoothed scale factors from drastically changing by bands. The described smoothing in time and bands of scale factors results in processed audio containing fewer undesirable perceptual artifacts.

Noise compensation

[b, t]. Для достижения этого результата можно использовать концепцию громкости частичных тонов, как определено Муром и Глазбергом ранее. Допустим, что кто-то способен получать измерение шума самого по себе и измерение аудио самого по себе. Пусть E _N[b, t] представляет накачку от шума, и пусть E _A[b, t] представляет накачку от аудио. Объединенная удельная громкость аудио и шума, в таком случае, задается посредствомIn many audio playback environments, there is background noise that interferes with the audio that the listener wants to listen to. For example, a listener in a moving automobile may be playing music through the integrated stereo system and noise from the engine and road may significantly alter the perception of the music. In particular, for parts of the spectrum in which the noise energy is significant relative to the energy of the music, the perceived loudness of the music decreases. If the noise level is high enough, the music is completely hidden. Regarding an aspect of the present invention, someone would like to select gain factors G [ b , t ] so that the specific volume of the processed audio in the presence of interfering noise is equal to the target specific volume

[ b , t ]. To achieve this result, you can use the concept of the volume of partial tones, as defined by Moore and Glazberg earlier. Suppose someone is able to receive a measurement of noise in itself and a measurement of audio in itself. Let E _N [ b , t ] represent pumping from noise, and let E _A [ b , t ] represent pumping from audio. The combined specific volume of audio and noise, in this case, is set by

where again Ψ {·} represents a nonlinear transformation from pump to specific volume. It can be assumed that the listener's hearing separates the combined specific volume between the specific volume of the partial tones of the audio and the specific volume of the partial tones of the noise in some way that preserves the combined specific volume:

Specific audio volume _{partials, N A [b, t]} , is a value that is desirable to control, and therefore it is necessary to calculate this value. The specific volume of partial noise tones can be approximately expressed as

where E _TN [ b , t ] is the masked threshold in the presence of noise, E _TQ [ b ] is the audibility threshold in silence on band b , and k is the exponent between zero and one.

Combining equations 31-33, we come to the expression for the specific volume of partial tones of audio:

It is noted that when the pumping of the audio is equal to the masked threshold noise value ( E _A [ b , t ] = E _TN [ b , t ]), the specific volume of the partial tones of the audio is equal to the volume of the signal at the threshold value in silence, which is the desired result. When the pumping of audio is much larger than that of noise, the second term in equation 34 disappears, and the specific volume of the audio is approximately equal to what it would be if there were no noise. In other words, while audio becomes much louder than noise, noise lurks behind the audio. The indicator k is chosen empirically in order to maximize compliance with the data on the volume of the tone in the noise as a function of the signal-to-noise ratio. Moore and Glasberg found that a value of k = 0.3 is appropriate. The masked threshold value of noise can be approximately expressed as a function of perception of noise itself:

where K [ b ] is a constant that increases in the low frequency bands. Thus, the specific volume of the partial audio tones defined by equation 34 can be represented abstractly as a function of the pump pumping audio and noise:

In this case, the modified gain factor solver can be used to calculate the gain G [ b , t ] so that the specific volume of the partial tones of the processed audio in the presence of noise is equal to the target specific volume:

Fig. 17 depicts the system of Fig. 7 with the original Gain Solver 106 replaced by the described noise compensation Gain Solver, 206, (note that the multiple vertical lines between the blocks representing the multiple filter bank bands have been replaced by a single line to simplify the circuit ) In addition, the figure depicts a measurement of noise pumping (Comb of analysis filters, 200, Pass filter 201, Pumping, 202, and Smoothing, 203, in some way corresponding to the operation of blocks 100, 101, 102 and 103), which is supplied to the new gain factor solver 206 along with pumping audio (from Smoothing, 103) and target specific volume (from SL Modification, 105).

[b, t] равной исходной удельной громкости аудио, N[b, t]. Другими словами, Модификация SL предусматривает независящий от частоты масштабный коэффициент α, масштабирующий удельную громкость звукового сигнала, при этом α=1. С помощью компоновки, такой как на фиг.17, коэффициенты усиления рассчитываются так, что воспринимаемый спектр громкости обработанного аудио в присутствии шума равен спектру громкости аудио в отсутствие шума. Дополнительно, любая одна или комбинация таковых из описанных ранее технологий для вычисления целевой удельной громкости в качестве функции исходной, в том числе VC, АРУ, DRC и DEQ, могут использоваться в соединении с системой модификации громкости с компенсацией шума.In its most basic operating mode, Modification SL, 105 of FIG. 17 may simply set a target specific volume

[ b , t ] equal to the original specific audio volume, N [ b , t ] . In other words, SL modification provides a frequency independent scaling factor α, the specific loudness scaling the audio signal, with α = 1. Using the arrangement, such as in FIG. 17, the gains are calculated so that the perceived loudness spectrum of the processed audio in the presence of noise is equal to the audio loudness spectrum in the absence of noise. Additionally, any one or a combination of those of the previously described technologies for calculating the target specific volume as a function of the original, including VC, AGC, DRC and DEQ, can be used in conjunction with a noise compensation volume modification system.

In a practical embodiment, the noise measurement can be obtained from a microphone located in or near the environment in which the audio will be played. Alternatively, a predefined set of noise pattern pumping that approximates the expected noise spectrum under various conditions can be used. For example, noise in the car interior can be pre-analyzed at various driving speeds, and then stored as a reference table of noise pumping depending on speed. The noise pumping supplied to the Gain Solver, 206, of FIG. 17, is then approximately expressed in this look-up table as the vehicle speed changes.

Approximation to specific volume

Although the disclosed invention works best when accurate specific volume measurements are used, some applications may require a coarser approximation in order to reduce computational complexity. With a suitable approximation, acceptable estimation and modification of perceived loudness can still be achieved. Such an approximation should try to preserve, at least in part, several key aspects of loudness perception. Firstly, the approximation should, at least roughly, record the change in sensitivity in the perception of volume depending on the frequency. In general terms, the approximation should display lower sensitivity at the lower and upper frequencies compared to the middle frequencies. Secondly, the approximation should demonstrate a nonlinear increase in volume depending on the signal level. More specifically, the increase in specific volume should be the fastest for low-level signals near the threshold of audibility, and then decrease to a constant growth rate as the signal level rises. Finally, the approximation should demonstrate the property of summing volume, meaning that for a constant signal level, the total volume (integration of the specific volume over frequency) increases as the signal bandwidth increases.

One way to reduce the complexity of calculating specific loudness, while still maintaining the desired properties of loudness perception, is to use a comb of filters with fewer bands, and in which the bands may not be evenly spaced along the step scale of critical bands. For example, a 5-band filter bank with strips distributed uniformly on a linear frequency scale, as opposed to strips 40 described previously. There are many technologies for the efficient implementation of such filter banks, for example, cosine modulated filter banks (PP Vaidyanthan, MultiRate Systems and Fitter Banks, 1993 Prentice Hall (P.P. Vaidyantan, Multistage filter systems and filters , 1993, Prentice Hall)) . As a general example, consider a comb filter with B bands, wherein each band is described by the impulse response h _b [n] in the time domain. In addition, suppose that the filter bank is an almost perfect reconstruction, bearing in mind that:

The frequency response of each band b can be characterized by a center frequency f _b and a bandwidth Δ f _b of the passband in Hertz. The equivalent center frequency and bandwidth in units of ERB, in this case, are set according to

If the number of B bands is relatively small, then the bandwidth e _b of the bandwidth of each band is likely to be greater than 1 ERB.

[b, n] может вычисляться подобно уравнениям 7 и 8, с вычислением сглаженной среднеквадратической мощности x _b [n], взвешенной частотной характеристикой пропускающего фильтра, дискретизированного на частоте f _b, и инверсией ширины полосы пропускания в единицах ERB:On the assumption that _{x b [n] = h b} [n] * x [n] represents an audio signal associated with the each lane smoothed pump signal

[ b , n ] can be calculated similarly to equations 7 and 8, with the calculation of the smoothed rms power x _b [ n ], the weighted frequency response of the pass filter, sampled at frequency f _b , and the inverse of the bandwidth in units of ERB:

Band b by pumping Weigh 1 / e _b effectively distributes the energy within such a band, evenly over all critical bands, referred to a group within it. Alternatively, someone could assign all the energy to a critical band whose center frequency is closest to the center frequency f _{b of the} band, but the energy distribution is uniformly the best approximation for most real sound signals.

[b, n], можно приступать к вычислению удельной громкости N[b, n], целевой удельной громкости

[b, n] и коэффициентов усиления G[b, n], как обсуждено выше, за исключением того, что здесь общее количество полос может быть гораздо меньшим, тем самым снижая сложность. К тому же, одной модификации необходимо производиться в вычислении общей громкости L[n]: удельная громкость должна взвешиваться количеством ERB в пределах полосы b, в то время как она суммируется по полосам:Pumped

[ b , n ], we can begin to calculate the specific volume N [ b , n ], the target specific volume

[ b , n ] and gain G [ b , n ], as discussed above, except that here the total number of bands can be much smaller, thereby reducing complexity. In addition, one modification must be made in calculating the total volume L [ n ]: the specific volume should be weighted by the number of ERBs within the band b, while it is summed over the bands:

In conclusion, the modified signal y [n] can be generated by summing each of the bandpass signals, weighted gain of the respective strips:

For simplicity of description, the above calculations are shown to be performed for each period n of the sample signal x [ n ]. In practice, however, the pump can be subjected to downsampling in time to a much lower frequency, and then all subsequent processing of the volume can be performed at this reduced frequency. When the gain coefficients are applied at the end, they can then be subjected to the interpolation by the upsampled before being applied to bandpass signals.

The above is only one example of a rough approximation to the specific volume, which is suitable for the disclosed invention. Other approximations are possible, and the invention is intended to cover the use of all such approximations.

Implementation

The invention may be implemented in hardware or software, or a combination of both (e.g., programmable logic arrays). Unless otherwise specified, the algorithms included as part of the invention are inherently irrelevant to any particular computer or other device. In particular, various general-purpose machines can be used with programs written in accordance with the doctrines given in the materials of this application, or it may be more convenient to design a more specialized device (for example, integrated circuits) to perform the required steps of the method. Thus, the invention can be implemented in one or more computer programs running in one or more programmable computer systems, each of which contains at least one processor, at least one data storage system (including volatile and non-volatile memory and / or storage elements), at least one input device or port, and at least one output device or port. The control program is applied to the input data to perform the functions described in the materials of this application, and generates output information. The output is fed to one or more output devices in a known manner.

Each such program can be implemented in any desired computer language (including machine, layout, or high-level procedural, logical, or object-oriented programming languages) for communicating with a computer system. In any case, the language may be a compiled or interpreted language.

Each such computer program is preferably stored on or loaded into storage media or device (e.g., solid state memory or media, or magnetic or optical media) readable by a general purpose or special purpose programmable computer to configure and control the computer when the storage media or devices are read a computer system to perform the procedures described in the materials of this application. The inventive system can also be considered as being implemented as a computer-readable storage medium configured by a computer program, where the storage medium configured in this way causes the computer system to operate in a special and predetermined manner to perform the functions described in the materials of this application.

A number of embodiments of the invention have been described. However, it will be understood that various modifications can be made without departing from the spirit and scope of the invention. For example, some of the steps described in the materials of this application may be independent of order and, thus, may be performed in order different from that described.

Claims (26)

1. A method for adjusting a specific characteristic of the loudness of an audio signal, in which a specific characteristic of loudness is either an approximation of the specific loudness, which is an indicator of perception loudness as a function of frequency and time, or an approximation of the specific loudness of partial tones, which is an indicator of the loudness of signal perception in the presence of a secondary interfering signal, as a function of frequency and time, consisting in the fact that
getting closer to the target specific loudness, getting frequency-dependent modification parameters used to modify the audio signal in order to reduce the difference between its specific loudness characteristic and approaching the target specific loudness, and
a) modification parameters apply to the audio signal to reduce the difference between its specific loudness characteristic and an approximation to the target specific loudness or
b) transmit or store the modification parameters and the audio signal for temporarily and spatially separated application of the modification parameters to the audio signal in order to reduce the difference between its specific loudness characteristic and approaching the target specific loudness. 2. The method according to claim 1, in which approaching the target specific volume is not a function of the audio signal. 3. The method according to claim 2, in which said modification or said receipt is that they take an approximation to the target specific volume from a source external to the method. 4. The method according to claim 2, in which said modification or said receipt includes processing that explicitly calculates the approach to the specific volume and / or the approximation to the specific volume of partial tones. 5. The method according to claim 2, in which said modification or said receipt includes processing that implicitly calculates an approach to specific volume and / or an approximation to specific volume of partial tones. 6. The method according to claim 5, in which the processing uses a look-up table so that the processing, by its nature, determines an approach to the specific volume and / or approach to the specific volume of partial tones. 7. The method according to claim 5, in which the approximation to the specific volume and / or the approximation to the specific volume of partial tones, by its nature, is determined by the mathematical expression reflecting a number of solutions used by the processing. 8. The method according to claims 2 to 7, in which the approximation to the target specific volume is independent of time and frequency. 9. The method according to claims 2 to 7, in which the approximation to the target specific volume is time-independent. 10. The method according to claim 1, in which the approximation to the target specific volume is a function of the sound signal or indicator of the sound signal. 11. The method according to claim 10, in which the indicator of the sound signal is approaching the specific volume of the sound signal. 12. The method of claim 10 or 11, wherein the function of the sound signal or measure of the sound signal is one or more scaling of the sound signal or measure of the sound signal. 13. A method for controlling a specific loudness characteristic of an audio signal, in which a specific loudness characteristic is either an approximation of specific loudness, which is an indicator of perception loudness as a function of frequency and time, or an approximation of specific loudness of partial tones, which is an indicator of loudness of perception of an audio signal in the presence of a secondary interfering signal as a function of frequency and time, consisting in the fact that
receive from the transmitted data or reproduce from the storage medium an audio signal and
a) frequency dependent modification parameters for modifying an audio signal modification parameters were derived from the approximation to the target specific loudness or
b) approaching a target specific loudness or presenting an approximation to such a target specific loudness, and
modify the audio signal in response to a) the received modification parameters, or b) the modification parameters obtained from approaching the target specific volume or its representation in order to reduce the difference between the specific characteristic of the volume of the audio signal and approaching the target specific volume. 14. The method according to item 13, in which approaching the target specific volume is not a function of the sound signal. 15. The method according to 14, in which the aforementioned modification or the aforementioned receipt is that take the approach to the target specific volume from a source external to the method. 16. The method of claim 14, wherein said modification or said preparation includes processing that explicitly calculates an approach to specific volume and / or an approach to specific volume of partial tones. 17. The method of claim 14, wherein said modification or said receipt includes processing that implicitly calculates an approach to specific volume and / or an approximation to specific volume of partial tones. 18. The method according to 17, in which the processing uses a look-up table so that the processing, by its nature, determines an approach to the specific volume and / or approach to the specific volume of partial tones. 19. The method according to 17, in which the approximation to the specific volume and / or the approximation to the specific volume of partial tones, by its nature, is determined by the mathematical expression reflecting a number of solutions used by the processing. 20. The method according to any one of paragraphs.14-19, in which the approximation to the target specific volume is independent of time and frequency. 21. The method according to any one of paragraphs.14-19, in which the approximation to the target specific volume is time-independent. 22. The method according to item 13, in which the approximation to the target specific volume is a function of the sound signal or indicator of the sound signal. 23. The method according to item 22, in which the indicator of the sound signal is approaching the specific volume of the sound signal. 24. The method according to item 22 or 23, in which the function of the audio signal or indicator of the audio signal is one or more scaling of the audio signal or indicator of the audio signal. 25. A device for controlling a specific characteristic of the volume of the audio signal, configured to implement all the steps of the method according to any one of claims 1 to 24. 26. A computer-readable medium containing a computer program recorded thereon, while the computer program, when executed by a computer, causes the computer to carry out all the steps of the method according to any one of claims 1 to 24.

Images